AI engineer - developer tools

About the role

This is a senior/staff-scope AI engineering role on a small team building Lighthouse AI, the intelligence engine for engineering leaders, and Clara, the company's agentic developer tooling initiative. You will own workstreams end to end: problem framing, data access, model choice, evaluation, rollout, iteration, and the product surface that exposes the output. The team needs someone who can bring classical ML discipline to the work while making practical decisions about LLMs, agents, and product integration. You will work closely with product, data, platform, and engineering stakeholders without a heavy process layer between you and the problem.

Responsibilities

• LLM-powered product features: build natural language query generation, summarization, and exploration tools that help users interact with engineering data directly.
• Model evaluation: design the evals, benchmarks, and review loops that tell the team whether an AI feature is actually useful, stable, and grounded.
• Causal and statistical modeling: work on the models behind engineering insights so the product can move from correlation to explanations leaders can act on.
• ML infrastructure: build and maintain pipelines for data processing, model inference, versioning, rollouts, and monitoring.
• Prompting and fine-tuning strategy: decide when prompting is enough, when retrieval or tool use is the right answer, and when fine-tuning or custom architectures are justified.
• Production quality: track regressions, latency, cost, and failure modes so AI features hold up under real customer usage.
• Product integration: work with product and engineering to ship AI directly into the user experience instead of treating it as a separate layer.
• Agentic workflows: help define how Clara should behave when AI systems interact with real software delivery workflows and live engineering context.

Requirements

• 5+ years of experience in AI, ML, or applied engineering, with at least one production system you owned materially end to end.
• Shipped ML or LLM features where evaluation, latency, cost, and data quality were real constraints.
• Worked on systems where a demo was not enough and the model had to keep working once real users started relying on it.
• Good judgment about when to use classical ML, LLMs, retrieval, tools, prompting, or fine-tuning.
• The ability to translate research ideas into systems that can ship and survive in production.
• Comfort working from ambiguous product goals and turning them into a technical plan with clear tradeoffs.
• The habit of instrumenting the system so you can tell whether it is helping, regressing, or drifting.
• Enough technical range to work across model design, infrastructure, and product integration without losing depth.

Qualifications

• Comfort working from ambiguous product goals and turning them into a technical plan with clear tradeoffs.
• The habit of instrumenting the system so you can tell whether it is helping, regressing, or drifting.
• Enough technical range to work across model design, infrastructure, and product integration without losing depth.

Skills

• Strong background in machine learning, natural language processing, and applied engineering.
• Experience with model evaluation, benchmarking, and review loops.
• Knowledge of causal and statistical modeling techniques relevant to engineering contexts.
• Expertise in building and maintaining machine learning pipelines.
• Ability to make tradeoffs between quality, latency, and cost.
• Experience with prompt engineering and fine-tuning strategies.
• Understanding of production quality metrics and monitoring.
• Proven ability to integrate AI features into product experiences.
• Experience with defining and implementing agentic workflows.

Benefits

• Competitive equity package.
• Hybrid work model with in-person time in the San Mateo office 3 days per week.
• Full-time employment status.
• Base salary range: $210K–$250K.

Pay

• Base salary range: $210K–$250K.

Schedule

• Hybrid work model with in-person time in the San Mateo office 3 days per week.