Senior AI Engineer
Confirmed on the employer's own hiring board on Nov 8, 2026. First seen by Alion on Nov 4, 2026. Raydar scores B on the Alion truth index.
About the company
Our client is a software company.
The role
Raydar is recruiting for this role on behalf of our client. Design and build the software layer that lets language models carry multi-step tasks to completion, covering tooling, context, state and feedback. The work runs from rapid prototyping through evaluation to hardening for reliable production use, with ownership at every stage. The role suits an engineer who wants to decide how much independent action an automated system should take.
What you'll do
- Build the scaffolding that gives language models tools, context and execution environments to finish multi-step tasks.
- Define how automated agents plan, act, track state, recover from errors and pass work back to people.
- Prototype new AI-driven features, test them on real user problems and move the strongest into production.
- Create evaluation suites that score task completion and decision quality, and turn production failures into repeatable test cases.
- Diagnose faults across the whole stack, from the instructions given to a model through its tool connections and workflow sequencing.
- Widen what automated systems can do on their own, setting explicit access limits and review points, and bringing in people when a workflow calls for it.
- Test new models and techniques for capability, reliability, latency and cost on real workflows.
- Build reusable tooling so new features are faster to ship and easier to operate.
Requirements
What we're looking for
- Strong software engineering skills and a track record of building agent systems that take actions and finish tasks.
- Working knowledge of how harness design, context handling and tool interfaces shape model output.
- Experience writing evals, reading execution traces and steadily fixing unreliable behavior.
- Speed in turning a concept into a functioning prototype, followed by sound choices about what needs hardening before release.
- Sound judgment about autonomy: when an agent may act alone, how to limit its actions, and when to involve a person.
- Curiosity about the product and ownership of results after launch.
- Background as a software engineer who has built, launched, run and fixed production software, rather than mainly consulting, client delivery or DevOps-only roles.
- Experience at an early-stage start-up.
- Around one to one and a half years of applied experience developing language model or agent software that people rely on, such as orchestration layers, function calling and evaluation pipelines.
- Track record of raising the quality, responsiveness or running expense of a production language model system through testing tools, log review, and performance and spending analysis.
Bonus points
- A computer science degree from a top 50 US university.
- Familiarity with agents that run for extended periods, including reliable execution, progress tracking, retry logic and recovery from faults.
- Exposure to agents that write code, operate web browsers or desktops, or coordinate several software tools and external interfaces.
- Creating test frameworks, simulated environments or utilities for reviewing and rerunning agent sessions.
- Delivering AI-enabled product features within a compact group while staying current as model abilities evolve.
Benefits
Compensation and benefits
- Base salary: USD 200,000 to 250,000 per year
- Equity
Location and work model
- San Francisco, CA, United States
- Hybrid, with Tuesdays remote
- Full-time