Experience level: 3–6-years - relevant

Location: 5-day on-site role in San Francisco

Client need engineers who have shipped LLM or agent features to production in TypeScript or Python. 

The best fits are engineers with 3–6 years of experience who have production LLM or agentic work at a Series A–D startup and are open to working on-site in SF.

Core responsibilities

• Own platform architecture for multimodal APIs, real-time event systems, AI orchestration and data infrastructure.
• Build production-grade agentic and conversational AI, including LLM integration, retrieval pipelines, tool orchestration and evaluations.
• Turn ambiguous product ideas into engineering roadmaps and lead technical initiatives across teams.
• Set service contracts, observability and reliability standards to support engineering growth from 30 to 300 engineers.
• Mentor senior engineers and integrate Claude Code, Cursor and other AI tools into everyday development.

Must have

• 8–15 years of software engineering experience and demonstrated Staff-level individual contributor impact.
• Core systems built and operated in production at scale; zero-to-one product delivery with end-to-end domain ownership.
• Strong system design, architecture review and current hands-on coding experience.
• At least two years of recent experience at Series A–D startups. Earlier Big Tech experience is acceptable.
• Daily, deep use of frontier AI tools such as Claude and Cursor, with specific examples of improved engineering output.
• Strong engineering fundamentals; Node.js, React and Python proficiency is preferred.
• BS in Computer Science or a STEM-related field.
• Willingness to work onsite in San Francisco five days per week, including a concrete relocation plan if needed.

Preferred experience

• Former technical founder or founding engineer.
• Conversational or voice AI, fintech, banking, or regulated financial services.
• TypeScript, JavaScript and Python; LLMs, agentic AI and RAG or retrieval pipelines.
• Event-driven architecture, distributed systems and cloud-native infrastructure.

Recruiter calibration

Assess architecture ownership, production impact and coding depth together. A strong candidate can explain design tradeoffs, reliability or latency decisions, customer impact, and what they personally implemented. Ask for measurable results rather than accepting team-level claims.
The listing flags repeated stints below two years and stagnant large-company careers without promotion roughly every two to three years. Clarify tenure patterns and career progression before submission. Prioritize builders over profiles limited to vendor configuration, implementation services, narrow maintenance scopes, or people management without current technical work.

Submission evidence

• Current resume and LinkedIn URL; startup stage and dates for relevant employers.
• Two or three examples of personally owned systems or products, including scale, outcomes and difficult architectural decisions.
• Current coding responsibilities and concrete AI-tool workflow examples.
• Location, onsite commitment, relocation timeline, base expectations, availability and job-search activity.
• Work authorization and employer-confirmed transfer or sponsorship requirements.