Staff AI Engineer

Lead AI engineering to bring LLM-powered intelligence to permitting workflows, with end-to-end ownership from problem framing to production evaluation.

Responsibilities

• Own the AI problem space by setting the technical direction for how Pulley applies LLMs across multiple product surfaces, from ambiguity to architecture to shipped, iterated outcomes
• Convert permitting documents, city regulations, and jurisdiction workflows into structured, reliable outputs, including extraction, classification, retrieval, and agentic workflows over document sources not designed for machine reading
• Define the company evaluation and observability standard: establish ground truth, measure quality and regressions, and implement systems that make “model improvement” measurable and default for LLM feature teams
• Use AI agents as a daily practice by directing, reviewing, and shipping agent-driven work at high velocity while maintaining a clear quality bar
• Make long-range technical bets that shape what Pulley can build next year, including model and architecture choices and decisions to build versus buy, with ownership of production consequences
• Multiply engineering impact by establishing reusable patterns for LLM feature development, mentoring senior engineers toward larger scope, and improving team speed through systems, standards, and abstractions

Requirements

• 8+ years of software engineering experience, with a substantial portion focused on production LLM or ML systems
• Track record owning a significant AI domain end-to-end, including problem identification, architecture, delivery, and production ownership, plus non-glamorous requirements such as data quality, evaluation design, cost and latency, and failure handling
• Hands-on, production experience with large language models, including prompting, retrieval-augmented generation, structured extraction, tool use, and agentic workflows, with judgment about when each approach is the wrong tool
• Experience designing evals and making LLM-powered features reliable in production
• Experience building with AI coding agents where agents performed substantial implementation under your direction (not limited to autocomplete)
• Ability to architect durable systems with pragmatic tradeoffs
• Experience mentoring engineers and/or setting technical direction that other engineers delivered against
• Based in the San Francisco Bay Area and willing to work in person 4 days a week

Technologies

• LLMs
• ML
• Retrieval-augmented generation
• TypeScript
• React
• Google Cloud

Compensation

• USD 300,000 - 350,000 per year
• Offers Equity

Employment Details

• Employment type: Full time
• Department: Engineering
• Location: San Francisco, CA (hybrid)

Who You Are

• Thrives in ambiguity, preferring to define the right problem over executing a prewritten spec
• Product-minded, focused on whether solutions solve customer needs and willing to talk to users
• Rigorous about what “working” means, prioritizing evals over demos and building measurement before building features
• Strong opinions about balancing quality and velocity using the right tools, abstractions, and processes
• Defaults to organizational-scale ownership, taking initiative to fix missing systems, processes, or gaps without needing permission

Nice to Have

• Experience with document understanding at scale (OCR, layout-aware parsing, or vision-language models over scanned PDFs, drawings, or forms)
• Experience fine-tuning models or building data pipelines to produce training and evaluation sets from real-world usage
• Experience in construction tech, govtech, proptech, or other domains where the hardest work involves messy real-world documents and processes
• Experience with modern full-stack development, including working in application code that puts AI features in front of users (TypeScript, React, Google Cloud)
• Experience serving as the most senior AI engineer in a domain, the escalation point when others need answers
• Startup experience at the stage where you helped build the team, not only the product