Aisi jobs

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><p> </p> <h1 class="text-node"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"><strong>Expression of Interest - Cyber and Autonomous Systems Team</strong> <strong>(CAST)</strong></span></h1> <p class="text-node"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">CAST is open to expressions of interest from talented cyber security experts and software engineers of all levels of seniority who want to work at the forefront of frontier AI security but do not see a role listed that matches their skills and experience. Our talent team will reach out should a role arise that may be a good match.</span></p> <h2 class="text-node"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"><strong>About the team</strong></span></h2> <p class="text-node"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">AI capabilities in cybersecurity and autonomy are advancing faster than at any point in history. Frontier models can now work through multi-step network intrusions, discover and exploit software vulnerabilities, and carry out long-horizon technical tasks with increasing independence. These are extraordinary tools for scientific and economic progress, but also have the potential for serious harm if misused or deployed without adequate oversight.</span><br><br><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">The team evaluates the capability of both frontier and open-weight AI models in cybersecurity, autonomy and AI R&D, ensuring the UK government and its partners have an accurate view of risks and capabilities. We use realistic cyber ranges and a large CTF suite for our evaluations, run pre-deployment testing of frontier models, and collaborate with our partners across UK government, frontier labs, and NCSC. </span><br><br><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">Your</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> day-to-day work </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">might include</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> building state</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">-</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">of</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">-</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">the</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">-</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">art</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938"> evaluations (e.g.,</span><a href="https://arxiv.org/pdf/2603.11214" target="_blank"> <span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">cyber ranges</span></a><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">),</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> running pre-deployment testing exercises</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938"> (e.g.,</span><a href="https://www.aisi.gov.uk/blog/our-evaluation-of-claude-mythos-previews-cyber-capabilities" target="_blank"> <span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">Claude</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> Mythos Preview</span></a><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">), </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">validating</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> novel model behaviours </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">(e.g., </span><a href="https://www.aisi.gov.uk/blog/cheating-behaviour-in-frontier-model-evaluations" target="_blank"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">models cheating in evaluations</span></a><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">), and answering timely research questions (e.g.</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">,</span> <a href="https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber" target="_blank"><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">the capabilities of o</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">pen</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">-</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">weight mode</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">ls</span></a><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">). </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">Working</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> alongside</span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938"> you would be</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90"> engineers and researchers who care </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">deeply </span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">about the impact of their work, take pride in their craft, and </span><span class="attr" data-user-id="24725b38-0d03-4aca-8ec9-b541e272e938">have a high level of autonomy</span><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">.</span></p> <p class="text-node"><span class="attr" data-user-id="f7681a72-9cbb-4b47-8afd-333602b36b90">If that sounds exciting to you, we'd love for you to apply!</span></p><div class="content-pay-transparency"><div class="pay-input"><div class="title">Salary</div><div class="pay-range"><span>£65,000</span><span class="divider">—</span><span>£145,000 GBP</span></div></div></div><div class="content-conclusion"><h2><strong><span data-contrast="none">What We Offer</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":0,"335559739":0}"> </span></h2> <p><strong><span data-contrast="none">Impact you couldn't have anywhere else</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":240,"335559739":240}"> </span></p> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804&am

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><p><strong>Expression of Interest - Red Team</strong><br>AISI's Red Team is open to expressions of interest from talented research engineers and scientists of all levels of seniority who want to work at the forefront of frontier AI safety and security but do not see a role listed that matches their skills and experience. Our talent team will reach out should a role arise that may be a good match.<br><br><strong>Team Description</strong><br>The Red Team conducts cutting edge research to identify, evaluate and stress test vulnerabilities in frontier AI systems, sharing our findings directly with leading AI companies, UK officials and allied governments to inform deployment, research and policy decisions.</p> <p>The Red Team is composed of three specialised sub teams, each tackling a distinct set of challenges. While each sub team has its own focus, we work closely together, sharing methodologies, tooling, research insights and findings across the wider Red Team. Many of our most impactful projects draw on expertise from across all three sub teams, and we actively encourage collaboration in working to tackle the most pressing challenges in frontier AI safety and security.<br><br><strong><em>Alignment</em> </strong><br>The Alignment sub-team focuses on detecting, evaluating and understanding misalignment in frontier AI systems. Our work centres on loss of control risks, including deceptive alignment, research sabotage and self-exfiltration attempts. We carry out novel research to develop techniques for finding misalignment, investigate how to attribute misaligned behaviour to more fundamental concerns such as instrumental convergence, and conduct pre and post deployment evaluations. We share our findings with frontier AI companies and with the UK and allied governments to inform deployment, research and policy making, and we work directly with safety teams at frontier labs to help improve their alignment training and monitoring methodologies.<br><br><strong><em>Misuse</em> </strong><br>The Misuse sub-team focuses on stress testing frontier AI safeguards for dangerous capabilities, researching novel attack vectors and developing advanced automated attack tooling. Our work probes the robustness of safeguards against real world threats, helping to identify where frontier systems may be vulnerable to exploitation and where defences need to be strengthened. We share our findings with frontier AI companies, including Anthropic, OpenAI and DeepMind, as well as with key UK officials and other governments, to inform their deployment, research and policy decisions and to support the development of more resilient safeguards across the frontier AI ecosystem.<br><br><strong><em>Control</em> </strong><br>The Control sub-team partners with leading frontier AI companies to stress test control measures designed to prevent AI systems from causing harm. Drawing on techniques from adversarial machine learning, we develop algorithms to uncover a wide range of failures in control measures and use these findings to assess and strengthen them. These partnerships allow us to directly influence vital control measures, while our position within government enables us to bring our understanding of the current state of control measures to wider government as critical deployment, research, and policy decisions are made.</p><div class="content-pay-transparency"><div class="pay-input"><div class="title">Salary</div><div class="pay-range"><span>£65,000</span><span class="divider">—</span><span>£145,000 GBP</span></div></div></div><div class="content-conclusion"><h2><strong><span data-contrast="none">What We Offer</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":0,"335559739":0}"> </span></h2> <p><strong><span data-contrast="none">Impact you couldn't have anywhere else</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":240,"335559739":240}"> </span></p> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1"><span data-contrast="none">Incredibly talented, mission-driven and supportive colleagues.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1"><span data-contrast="none">Direct influence on how frontier AI is governed and deployed globally.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="3" data-aria-level="1"><span data-contrast="none">Work with the Prime Minister’s AI Advisor and leading AI companies.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="4" data-aria-level="1"><span data-contrast="none">Opportunity to shape the first & best-resourced public-interest research team focused on AI security.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <p><strong><span data-contrast="none">Resources & access</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":240,"335559739":240}"> </span></p> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685&

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><div class="vac_display_field"> <div class="vac_display_field_value"> <div class="vac_display_field"> <div class="vac_display_field_value"> <h2><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW212922402 BCX8" data-ccp-parastyle="heading 2">The deadline for applying to this role is </span><span class="NormalTextRun CommentStart CommentHighlightPipeRest CommentHighlightRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2">6</span></span><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun Superscript CommentHighlightRest SCXW212922402 BCX8" data-fontsize="12" data-ccp-parastyle="heading 2">th</span></span><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun CommentHighlightRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2"> September 2026</span><span class="NormalTextRun CommentHighlightPipeRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2">, end of day, anywhere on Earth.</span></span><span class="EOP Selected SCXW212922402 BCX8" data-ccp-props="{"134245418":true,"134245529":true,"335557856":16777215,"335559738":0,"335559739":0}"> </span></h2> <h2>About the team</h2> <p><span data-contrast="none">AISI's Chem Bio (CB) team conducts technical research to assess evolving AI capabilities related to science R&D and CB misuse, and the effectiveness of technical safeguards that might mitigate risks arising from those capabilities.</span><span data-ccp-props="{"335557856":16777215,"335559738":240,"335559739":240}"> </span></p> <p><span data-contrast="none">The goal of our research is to inform critical decisions on security, opportunities, policy, and risk mitigation made by governments and AI developers.</span><span data-ccp-props="{"335557856":16777215,"335559738":240,"335559739":240}"> </span></p> <p><span data-contrast="none">We're a close-knit, unusually interdisciplinary team—made up of machine learning researchers and engineers, software engineers, virologists and bacteriologists, behavioural research scientists, biosecurity experts, long-standing CB policy specialists and talented generalists—who work closely with other technical and policy teams across government.</span><span data-ccp-props="{"335557856":16777215,"335559738":240,"335559739":240}"> </span></p> <p><span data-contrast="auto">Over the next twelve months, CB will hugely scale the range and complexity of the evaluations and research programmes it carries out, and engage more deeply with partners in major AI labs, the wider biotech and pharma ecosystem and security services than it ever has before.</span><span data-ccp-props="{"335557856":16777215,"335559738":240,"335559739":240}"> </span></p> <p> </p> <h2>About the role</h2> <p>AI capabilities in the life sciences are advancing faster than at any point in history. <span class="NormalTextRun CommentStart SCXW35161139 BCX8">Foundation models can now design novel proteins</span><span class="NormalTextRun SCXW35161139 BCX8"> and</span><span class="NormalTextRun SCXW35161139 BCX8"> interpret genomic sequences. Specialised biological models can </span><span class="NormalTextRun SCXW35161139 BCX8">both </span><span class="NormalTextRun SCXW35161139 BCX8">identify</span><span class="NormalTextRun SCXW35161139 BCX8"> drug targets and design the compound to target them.</span> These are extraordinary tools for scientific progress but also have the potential for harm if misused.  </p> <p><span class="TextRun SCXW106540103 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW106540103 BCX8">This role is for a technical researcher who can contribute strong ML and computational biology </span><span class="NormalTextRun SCXW106540103 BCX8">expertise</span><span class="NormalTextRun SCXW106540103 BCX8"> to that mission. You will sit within a group of research scientists, subject matter experts and engineers, leading empirical research into the risk-relevant capabilities of specialised biological models, including biomolecular structure and generative-design systems. You will translate ambiguous questions about what these models could enable into rigorous research questions and experimental designs, assess whether in-silico performance translates into meaningful experimental outcomes, and investigate whether technical safeguards can reliably limit potentially dangerous capabilities. It is a role at the interface of machine learning, computational </span><span class="NormalTextRun SCXW106540103 BCX8">biology</span><span class="NormalTextRun SCXW106540103 BCX8"> and biosecurity: shaping which capabilities we investigate, how we measure their real-world significance, and how we translate our findings into decisions by government and other trusted partners.</span></span><span class="EOP Selected SCXW106540103 BCX8" data-ccp-props="{"335557856":16777215,"335559739":0}"> </span></p> <p> </p> </div> </div> <div class="vac_display_field"> <h2><span class="TextRun SCXW166520122 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW166520122 BCX8" data-ccp-parastyle="heading 2">What you will own</span></span><span class="EOP Selected SCXW166520122 BCX8" data-ccp-props="{"134245418":true,"134245529":true,"335557856":16777215,"335559738":0,"335559739":0}"> </span></h2> <div class="vac_display_field_value"> <ul> <li><strong>Evaluate the risk-relevant capabilities of specialised biological models: </strong>Translate important but ambiguous questions about the capabilities of state-of-the-art biological models (including biomolecular structure and generative design models) into measurable research questions and experimental designs. </li> <li><strong>Connect computational and experimental evidence:</strong> Use published experimental results, biological datasets, expert review and, where appropriate, external wet-lab collaborations to assess whether in-silico performance translates into experimentally relevant outcomes. </li> <li><strong>Identify feasibility and effectiveness of technical safeguards: </strong>Lead research into the feasibility and effectiveness of technical safeguards for specialised biological models, including access controls, model-level interventions, monitoring, detection and capability-limiting approaches. </li> <li><strong>Track the technical frontier:</strong> Identify important developments in biological AI and determine which new models, methods or capabilities AISI should investigate. Help shape the team’s research agenda as the field evolves. </li> <li><span class="TextRun SCXW50782935 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW50782935 BCX8"><strong>Communicate to decision-makers:</strong> </span></span><span class="TextRun SCXW50782935 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW50782935 BCX8">Produce clear technical reports, </span><span class="NormalTextRun SCXW50782935 BCX8">briefings</span><span class="NormalTextRun SCXW50782935 BCX8"> and recommendations for senior decision-makers within government and other tru

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><div class="vac_display_field"> <div class="vac_display_field_value"> <div class="vac_display_field"> <div class="vac_display_field_value"> <h2><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW212922402 BCX8" data-ccp-parastyle="heading 2">The deadline for applying to this role is </span><span class="NormalTextRun CommentStart CommentHighlightPipeRest CommentHighlightRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2">6</span></span><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun Superscript CommentHighlightRest SCXW212922402 BCX8" data-fontsize="12" data-ccp-parastyle="heading 2">th</span></span><span class="TextRun SCXW212922402 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun CommentHighlightRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2"> September 2026</span><span class="NormalTextRun CommentHighlightPipeRest SCXW212922402 BCX8" data-ccp-parastyle="heading 2">, end of day, anywhere on Earth.</span></span><span class="EOP Selected SCXW212922402 BCX8" data-ccp-props="{"134245418":true,"134245529":true,"335557856":16777215,"335559738":0,"335559739":0}"> </span></h2> <h2>About the team</h2> <p>AISI's Chem Bio (CB) team conducts technical research to assess evolving AI capabilities related to science R&D and CB misuse, and the effectiveness of technical safeguards that might mitigate risks arising from those capabilities. </p> <p>The goal of our research is to inform critical decisions on security, opportunities, policy, and risk mitigation made by governments and AI developers. </p> <p>We're a close-knit, unusually interdisciplinary team—made up of machine learning researchers and engineers, software engineers, virologists and bacteriologists, behavioural research scientists, biosecurity experts, long-standing CB policy specialists and talented generalists—who work closely with other technical and policy teams across government. </p> <p>Over the next twelve months CB will hugely scale the range and complexity of the evaluations and research programmes it carries out, and engage more deeply with partners in major AI labs, the wider biotech and pharma ecosystem and security services than it ever has before. </p> <p> </p> <h2>About the role</h2> <p>AI capabilities in the life sciences are advancing faster than at any point in history. Foundation models can now design novel proteins and interpret genomic sequences. Specialised biological models can both identify drug targets and design the compound to target them. These are extraordinary tools for scientific progress, but also have the potential for harm if misused.  </p> <p>This role is for a senior technical expert who can bring deep virology and biosecurity judgement to that mission. You will sit directly alongside the CB team’s researchers, helping design and interpret evaluations of how frontier and specialised AI systems perform on biologically relevant tasks, with a particular focus on pathogens and virology-relevant capabilities. You will help translate ambiguous biological risk questions into rigorous empirical evaluations, identify where model outputs are scientifically meaningful versus merely plausible-sounding, and advise on the mitigations, safeguards and evidence standards needed before those results can inform policy. It is a role at the interface of virology, AI evaluation, and biosecurity: shaping what we test, how we test it, and how we explain the significance of our findings to government and external partners. </p> <p> </p> </div> </div> <div class="vac_display_field"> <h2><span class="TextRun SCXW166520122 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW166520122 BCX8" data-ccp-parastyle="heading 2">What you will own</span></span><span class="EOP Selected SCXW166520122 BCX8" data-ccp-props="{"134245418":true,"134245529":true,"335557856":16777215,"335559738":0,"335559739":0}"> </span></h2> <div class="vac_display_field_value"> <ul> <li><span class="TextRun SCXW264198817 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW264198817 BCX8"><strong>Design risk-relevant biology evaluations</strong>: </span></span><span class="TextRun SCXW264198817 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW264198817 BCX8">Design, develop and execute evaluations that test the frontier biological capabilities and risk-relevant behaviours of frontier and specialised AI models, with a particular focus on virology-relevant tasks.</span></span> </li> <li><strong><span class="TextRun SCXW254632164 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW254632164 BCX8">Translate research </span><span class="NormalTextRun SCXW254632164 BCX8">objectives</span><span class="NormalTextRun SCXW254632164 BCX8"> into rigorous evidence: </span></span></strong><span class="TextRun SCXW254632164 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW254632164 BCX8">Help turn ambiguous biological risk questions into rigorous empirical claims, </span><span class="NormalTextRun SCXW254632164 BCX8">identifying</span><span class="NormalTextRun SCXW254632164 BCX8"> the right evaluation designs, evidence standards, scoring </span><span class="NormalTextRun SCXW254632164 BCX8">rubrics</span><span class="NormalTextRun SCXW254632164 BCX8"> and expert review processes needed to support high-confidence conclusions.</span></span><span class="EOP Selected SCXW254632164 BCX8" data-ccp-props="{"335557856":16777215,"335559739":0}"> </span> </li> <li><span class="TextRun SCXW177002648 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW177002648 BCX8"><strong>Interpret evaluation results responsibly:</strong> </span></span><span class="TextRun SCXW177002648 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW177002648 BCX8">Analyse and interpret evaluation findings, including their limitations, uncertainty, false-positive and false-negative risks, and implications for capabilities and risk. </span></span><span class="EOP Selected SCXW177002648 BCX8" data-ccp-props="{"335557856":16777215,"335559739":0}"> </span></li> <li><span class="TextRun SCXW35222914 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW35222914 BCX8"><strong>Communicate to decision-makers:</strong> </span></span><span class="TextRun SCXW35222914 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW35222914 BCX8">Produce clear technical reports, </span><span class="NormalTextRun SCXW35222914 BCX8">briefings</span><span class="NormalTextRun SCXW35222914 BCX8"> and recommendations for senior decision-makers within government and other trusted partners, translating complex scientific and AI-safety evidence into actionable conclusions.</span></span><span class="EOP Selected SCXW35222914 BCX8" data-ccp-props="{"335557856":16777215,"335559739":0}"> </span></li> <li><span class="TextRun SCXW223420276 BCX8" lang="EN-GB" data-contrast="none"><span class=&qu

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><div class="vac_display_field"> <div class="vac_display_field_value"> <div class="vac_display_field"> <div class="vac_display_field_value"> <h2><strong>About the Team</strong></h2> <p>AISI's Chem Bio (CB) team conducts technical research to assess evolving AI capabilities related to science R&D and CB misuse, and the effectiveness of technical safeguards that might mitigate risks arising from those capabilities.</p> <p>The goal of our research is to inform critical decisions on security, opportunities, policy, and risk mitigation made by governments and AI developers.</p> <p>We're a close-knit, unusually interdisciplinary team—made up of machine learning researchers and engineers, software engineers, virologists and bacteriologists, behavioural research scientists, biosecurity experts, long-standing CB policy specialists and talented generalists—who work closely with other technical and policy teams across government.</p> <h2><strong>Role Responsibilities</strong></h2> <p>We are building a dedicated engineering function within the CB team — a small team that owns the shared platform, tooling, and infrastructure that our research projects depend on. This role is a senior individual contributor within that function. The successful candidate will:</p> <ul> <li>Design and build reliable systems — architect and deliver core platform components that enable researchers to run experiments and ship results quickly, including LLM-based agent systems, evaluation frameworks, and supporting infrastructure.</li> <li>Work closely with researchers — translate varied research needs across general agents, science agents, chemical and biological models and other CB workstreams into well-engineered, maintainable systems.</li> <li>Own and deliver significant engineering work — take responsibility for substantial technical projects and shared components, driving them from design through implementation, iteration, and maintenance.</li> <li>Write high-quality production code — build scalable, robust, and maintainable software, primarily in Python, with strong engineering discipline around testing, documentation, and observability.</li> <li>Contribute to engineering standards and practices — help improve code quality, development workflows, and shared engineering approaches within a growing engineering team embedded in a research-heavy environment.</li> <li>Collaborate across AISI — work with our Core Technology and engineering teams to share tooling, align on infrastructure, and avoid duplication.</li> <li>Support technical direction — contribute to architectural decisions, identify opportunities to strengthen shared infrastructure, and help make pragmatic technical trade-offs.</li> <li>Mentor others through technical leadership — support other engineers through collaboration, code review, and knowledge sharing, without formal line-management responsibility.</li> </ul> <h2><strong>Role Requirements</strong></h2> <p>We are looking for the following skills, experience and attitudes, but a successful candidate will not necessarily need to meet all these criteria. We can be flexible in shaping the role and salary to your background, expertise, and level of experience.</p> <ul> <li>Significant experience writing production-level Python code that is scalable, robust, and easy to maintain, with a track record of delivering substantial technical work.</li> <li>Strong infrastructure and platform skills — experience with cloud environments (AWS), container orchestration (Kubernetes), and job scheduling (Slurm or similar).</li> <li>Experience owning technical projects or major systems as a senior individual contributor. You do not need formal management experience for this role.</li> <li>Familiarity with the AI/ML ecosystem: OpenAI-compatible APIs, PyTorch, and the tooling around LLM-based systems.</li> <li>Ability to work across multiple teams, understanding varied research needs and delivering reliable engineering solutions.</li> <li>A sense of mission, urgency, and responsibility for success: Demonstrated ability to solve challenging problems, implement solutions efficiently and acquire any missing knowledge necessary to get the job done. Motivated to build software with direct policy impact.</li> <li>Strong communication skills, with the ability to work effectively in a dynamic, multidisciplinary team environment.</li> </ul> <p><strong>Strong candidates may also have:</strong></p> <ul> <li>Experience working in a research environment — you don't need to be a researcher, but you should be comfortable working alongside them and translating ambiguous research requirements into engineering plans.</li> <li>Familiarity with computational biology tools, workflows, or datasets.</li> <li>Experience building or maintaining internal developer platforms, shared libraries, or CI/CD systems for technical teams.</li> </ul> <p>Please note that this is a reserved post. We can only consider applications from UK nationals (including dual nationals who hold British citizenship). Appointment is conditional on successfully completing UK Government SC clearance. Prior clearance is not required—we will sponsor and support you. You should normally have been resident in the UK for the past 5 years. You may also be required to undergo Developed Vetting (DV). DV typically requires a longer period of UK residency (around 10 years). Employment is conditional on obtaining and maintaining the required clearance(s). More detail on clearance eligibility can be found on the UK Government website: <a href="https://www.gov.uk/government/publications/united-kingdom-security-vetting-clearance-levels/national-security-vetting-clearance-levels">National security vetting: clearance levels - GOV.UK.</a></p> <p><strong>Other core requirements:</strong></p> <ul> <li>You should be able to spend at least 9 days per fortnight working with us.</li> <li>You should be willing to work from our office in London (Whitehall) at least 3 days/week.</li> <li>You should be UK-based.</li> </ul> </div> </div> </div> </div> <div class="vac_display_field"> </div><div class="content-conclusion"><h2><strong><span data-contrast="none">What We Offer</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":0,"335559739":0}"> </span></h2> <p><strong><span data-contrast="none">Impact you couldn't have anywhere else</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":240,"335559739":240}"> </span></p> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1"><span data-contrast="none">Incredibly talented, mission-driven and supportive colleagues.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1"><span data-contrast="none">Direct influence on how frontier AI is governed and deploye

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><div class="vac_display_field"> <div class="vac_display_field_value"> <div class="vac_display_field"> <div class="vac_display_field_value"> <h2>Role Description</h2> <p><span class="TextRun SCXW156930224 BCX0" lang="EN-US" data-contrast="auto"><span class="NormalTextRun SCXW156930224 BCX0">The AI Security Institute's Research Unit is looking for motivated and talented Software Engineers to join AISI's Core Technology Team. We are looking for exceptional candidates at all experience levels, from junior through to senior or staff, to work</span><span class="NormalTextRun SCXW156930224 BCX0"> in small teams on a range of critical research-oriented software and infrastructure.</span></span></p> <p>In this role, you’ll work with cutting-edge technologies on research problems with real-world impact, and receive mentorship and coaching from your manager and the technical leads on your team. You'll also regularly interact with world-famous researchers and other incredible staff (including alumni from Anthropic, DeepMind, OpenAI, Google, Apple, and professors from Oxford and Cambridge).</p> <h3>About our Core Technology Team</h3> <p style="font-weight: 400;">AISI Core Technology Team comprises several teams building tools and infrastructure used across all of our research work. This includes projects like Inspect (our open-source evaluation framework), systems for running evaluations at scale, and hosting frontier open-weights models for evaluations or human studies.</p> <p style="font-weight: 400;">As a software engineer on one of these teams, you might:</p> <ul> <li>Add a feature to one of our “sandbox plugins” for Inspect, supporting advanced agentic evals with safe mechanisms to let models write and execute arbitrary code</li> <li style="font-weight: 400;">Implement support for a new class of open-weights model on our model hosting platform</li> <li style="font-weight: 400;">Support a research team with designing custom infrastructure for a new research project</li> <li style="font-weight: 400;">Collaborate with the open-source community on a feature in Inspect or Inspect's plugin ecosystem</li> <li style="font-weight: 400;">Assist with an evaluation testing exercise of a frontier AI model, debugging issues that appear in Inspect from API changes in the lab's latest SDK</li> </ul> </div> </div> <div class="vac_display_field"> <h3>Person Specification</h3> <div class="vac_display_field_value"> <p>You may be a good fit if you have <strong>some</strong> of the following skills, experience and attitudes:</p> <ul> <li>Writing production quality code at fast pace.</li> <li>Designing, shipping, and maintaining complex tech products.</li> <li>Improving technical standards across a team, through mentoring and feedback.</li> <li>Strong written and verbal communication skills.</li> <li>Experience working with a world-class research team comprised of both scientists and engineers (e.g. in a top-3 lab).</li> <li>Python experience, including understanding the intricacies of the language, the good vs. bad Pythonic ways of doing things and much of the wider ecosystem/tooling.</li> <li>Experience building and maintaining systems on AWS or other cloud providers using infrastructure-as-code.</li> </ul> <p>Motivated candidates are encouraged to apply even if you don't meet all the above criteria.</p> <h3>Required Experience</h3> <p>We select based on skills and experience regarding the following areas:</p> <ul> <li>Writing production quality code</li> <li>Writing code efficiently</li> <li>Python</li> <li>Written communication</li> <li>Verbal communication</li> <li>Teamwork</li> <li>Interpersonal skills</li> <li>Tackling challenging problems</li> </ul> <h3>Desired Experience</h3> <p>We additionally may factor in experience with particular areas like:</p> <ul> <li>Expertise in Cloud Infrastructure or Dev Ops (AWS, Azure, Kubernetes, Terraform, CDK, Docker, etc.)</li> <li>Cybersecurity expertise</li> <li>ML Ops (vLLM, agent frameworks, fine-tuning, RAG systems, etc.)</li> </ul> </div> </div> </div> </div> <div class="vac_display_field"> <div class="vac_display_field_value"> <p> </p> </div> </div><div class="content-conclusion"><h2><strong><span data-contrast="none">What We Offer</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":0,"335559739":0}"> </span></h2> <p><strong><span data-contrast="none">Impact you couldn't have anywhere else</span></strong><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559685":0,"335559738":240,"335559739":240}"> </span></p> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1"><span data-contrast="none">Incredibly talented, mission-driven and supportive colleagues.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1"><span data-contrast="none">Direct influence on how frontier AI is governed and deployed globally.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="3" data-aria-level="1"><span data-contrast="none">Work with the Prime Minister’s AI Advisor and leading AI companies.</span><span data-ccp-props="{"134233117":false,"134233118":false,"335557856":16777215,"335559738":0,"335559739":0}"> </span></li> </ul> <ul> <li data-leveltext="" data-font="Symbol" data-listid="52" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel&

<div class="content-intro"><h2>About the AI Security Institute</h2> <p>The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.</p> <p>We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</p></div><div class="vac_display_field"> <div class="vac_display_field_value"> <div class="vac_display_field"> <div class="vac_display_field_value"> <h2><strong>The deadline for applying to this role is Sunday 30th August 2026, end of day, anywhere on Earth.</strong></h2> <h2><strong>About the role</strong></h2> <p><span data-contrast="none">AISI aims to turn progress on important research questions in AI security into real insight and impact for our partners and the public. This requires running research programmes that actually deliver, and making these programmes rigorous and fast is its own discipline. The AI Security Institute is hiring a Technical Programme Manager for our Cyber and Autonomous Systems team to do exactly this.</span><span data-ccp-props="{"335551550":1,"335551620":1}"> </span></p> <p><span data-contrast="none">Over the next twelve months, this team will need to move faster, deliver more complex research programmes and evaluations, and engage more deeply with partners in major AI labs and across government than ever before.</span><span data-ccp-props="{}"> </span></p> <p><span data-contrast="none">Technical Programme Managers make this work possible. You will work directly with the team's researchers and engineers to turn vast research questions into tractable programmes. You will be held responsible for ensuring our research stands up to rigorous empirical standards to inform policy, recommending which technical work needs to happen, and untangling knotty dependencies across workstreams.</span><span data-ccp-props="{}"> </span></p> <p> </p> <h2><strong>About the team</strong></h2> <p><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">AI capabilities in cybersecurity and autonomy are advancing fast. Frontier models can now work through multi-step cyberattacks, discover and exploit software vulnerabilities, and work independently for days with increasing reliability. These are extraordinary tools for scientific and economic progress but also have the potential for serious harm if misused or deployed without adequate oversight.</span></span><span class="LineBreakBlob BlobObject DragDrop SCXW134974837 BCX8"><span class="SCXW134974837 BCX8"> </span><br class="SCXW134974837 BCX8"></span><span class="LineBreakBlob BlobObject DragDrop SCXW134974837 BCX8"><span class="SCXW134974837 BCX8"> </span><br class="SCXW134974837 BCX8"></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">The AI Security Institute's Cyber and Autonomous Systems Team (CAST) exists to evaluate the capability of both frontier and open-weight AI models in cybersecurity, autonomy, and AI R&D, ensuring the UK government and its partners have an accurate view of risks and capabilities. We </span></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">continuously build</span></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8"> out</span><span class="NormalTextRun CommentStart SCXW134974837 BCX8"> </span></span><a class="Hyperlink SCXW134974837 BCX8" href="https://arxiv.org/pdf/2603.11214" target="_blank"><span class="TextRun Underlined SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8" data-ccp-charstyle="Hyperlink">more realistic cyber ranges</span></span></a><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8"> </span><span class="NormalTextRun SCXW134974837 BCX8">and</span></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8"> a CTF suite</span></span><span class="TrackChangeTextInsertion TrackedChange SCXW134974837 BCX8"><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">;</span></span></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8"> run pre-deployment testing of frontier models</span></span><span class="TrackChangeTextInsertion TrackedChange SCXW134974837 BCX8"><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">; </span></span></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">and collaborate with our partners across the UK and US governments, frontier AI labs, and the National Cyber Security Centre.</span></span><span class="LineBreakBlob BlobObject DragDrop SCXW134974837 BCX8"><span class="SCXW134974837 BCX8"> </span><br class="SCXW134974837 BCX8"></span><span class="LineBreakBlob BlobObject DragDrop SCXW134974837 BCX8"><span class="SCXW134974837 BCX8"> </span><br class="SCXW134974837 BCX8"></span><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8">We provide independent, empirical insight into the cyber capabilities of frontier and open-weight models, </span><span class="NormalTextRun SCXW134974837 BCX8">leveraging</span><span class="NormalTextRun SCXW134974837 BCX8"> our unique position within government and our partnerships with frontier labs</span><span class="NormalTextRun SCXW134974837 BCX8">. K</span><span class="NormalTextRun SCXW134974837 BCX8">ey examples </span><span class="NormalTextRun SCXW134974837 BCX8">are</span><span class="NormalTextRun SCXW134974837 BCX8"> our evaluation of </span></span><a class="Hyperlink SCXW134974837 BCX8" href="https://www.aisi.gov.uk/blog/our-evaluation-of-claude-mythos-previews-cyber-capabilities" target="_blank"><span class="TextRun Underlined SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8" data-ccp-charstyle="Hyperlink">Anthropic's Mythos Preview</span></span></a><span class="TextRun SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8"> </span><span class="NormalTextRun SCXW134974837 BCX8">and our research </span></span><a class="Hyperlink SCXW134974837 BCX8" href="https://www.aisi.gov.uk/blog/how-far-behind-the-frontier-are-leading-open-weight-models-on-cyber" target="_blank"><span class="TextRun Underlined SCXW134974837 BCX8" lang="EN-GB" data-contrast="none"><span class="NormalTextRun SCXW134974837 BCX8" data-ccp-charstyle="Hyperlink">measuring the gap between leading open-weight models and the frontier on cyber capabilities.</span></span></a><span class="EOP Selected SCXW134974837 BCX8" data-ccp-props="{}"> </span></p>