Research Engineer (AIR)

Department: Engineering

Location: San Francisco

Compensation: $200K – $350K • Offers Equity

Employment Type: FullTime

About AIUC

Agents are being deployed into hospitals, banks, and governments today. Soon there will be more agent-to-agent interactions than human-to-human interactions. The world needs a gold standard for AI safety, security, and reliability. That standard is AIUC-1, and building it is our mission.

Our mission is to underwrite superintelligence. We evaluate, certify, and insure AI agents, giving high-growth AI companies the credibility to earn their customers' trust. Over time, we will do the same for foundation models and data centers.

Our team comes from Anthropic, McKinsey, METR, Weights & Biases, Perplexity, and the Thiel Fellowship. We have raised $15M led by Nat Friedman, Emergence, and Terrain, with backing from Anthropic co-founder Ben Mann and ex-CISOs at Google Cloud and MongoDB.

The role

As an AI Researcher at AIUC, you will develop and expand the evaluation methods that sit at the heart of our work. You will identify the most pressing problems in our evaluation stack, scope and lead projects to address them, and push the frontier of what rigorous AI assessment looks like.

Your work spans three horizons. On the product side, you'll improve our scale and accuracy by building better LLM judges, tightening our pipelines, making our evaluations faster and more reliable. On the more pure research side, you'll deepen the quality of what we evaluate. This means designing new attack vectors, implementing techniques from the latest research, and building agentic automations that extend our capabilities. And on the longer horizon, you'll take on moonshot projects: things like fully dynamic attackers, self-expanding libraries of attacks, and novel approaches to evaluation that don't yet exist.

From the outset you will:

• Identify and scope the highest-leverage problems in our evaluation system, then lead projects end-to-end to address them.

• Build novel approaches to AI evaluation by implementing research papers, replicating attack techniques, and experimenting with new methods.

• Lead and coordinate research teams, managing complex multi-person projects with clear ownership and delivery.

• Communicate findings internally and externally through technical blog posts, papers, and direct engagement with client partners.

• Feed insights back to the product, shaping our roadmap based on what you learn on the frontier.

What we're looking for

We're looking for researchers who combine technical depth with strong product instincts.

You'll need to have:

• Deep AI industry knowledge: You understand the techniques behind language model agents, LLM pipelines, and agentic systems. You know the tools, the players, and the direction the field is moving.

• Strong technical foundation: Solid coding ability and statistical analysis skills. You're comfortable implementing research and building systems that others depend on.

• Clear communication: You can collaborate effectively across engineering, product, and client-facing teams. You translate complex technical work into language that lands with any audience.

• Research leadership: You've managed complex projects end-to-end and can coordinate across teams of researchers with clarity and conviction.

• Culture fit: You are genuinely motivated by AIUC's mission. You thrive in a high-intensity environments and bring openness, honesty, and emotional maturity to the team.

Nice to have:

• Hands-on experience with AI red teaming, evals, post-training, or alignment research.

• Prior experience in a fast-moving startup environment.

• Published research or technical writing in AI safety, security, or related fields.

Best fit: A research background combined with product instincts and experience shipping real systems not just writing papers about them.

Our values

Strong Back

We hold ourselves to the highest standards because underwriting superintelligence demands nothing less. Trust is our business, and we earn it through crisp thinking, radical ownership, and delivering on every promise we make.

Fast Feet

We move fast by shipping early, deciding decisively, and cutting what doesn't deserve full effort. Speed comes from preparation, experimentation, and always knowing who owns what and by when.

Eyes Up

We stay sharp on changes in AI and how they impact our product, our customers, and the world. AI fluency is not optional: we use it, experiment with it, and integrate it into our daily workflows.

Open Heart

We show up honestly, name what's real, and lean into the hard conversations rather than manage around them. Real connection comes from sharing your experience, not controlling the outcome.