AI Engineer

Feltsense; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Feltsense is hiring a Head of Engineering to own the autonomous multi-agent system that runs from idea to live ads to launched software. You will own the architecture, set technical direction, and lead the engineering team at an early stage.

You will ship production code, optimize agent runtimes, and drive improvements in AI behavior, prompts, and system reliability, latency, and cost in production. This role demands hands-on leadership and deep backend expertise in a fast-growing startup.

Qualifications

• 4+ years of professional software engineering experience at an early-stage company.
• Hands-on experience with multi-agent systems and orchestration, shipped to production.
• Strong backend engineering skills: designing, building, and maintaining APIs and data flows.
• Experience with prompt engineering, model configuration, structured outputs, embeddings, retrieval, and evaluation.
• Daily, hands-on use of AI coding agents; you plan work, hand it to agents, review what they produce, and know when to verify instead of trust.

Responsibilities

• Own the system architecture and make the technical calls; move the team through decisions rather than toward consensus.
• Break a large technical problem into the first wins that build momentum, and sequence the roadmap around them.
• Ship production code yourself, especially on the agent runtime and orchestration.
• Debug and improve AI behavior in production; drive prompt quality, structured outputs, retries, and fallbacks.

Skills

Engineering leadership

Multi-agent systems

Backend API design

Prompt engineering

AI tooling

TypeScript

Node.js

Vercel AI SDK

Claude Agent SDK

agent runtimes

AI safety/observability

Tools

Anthropic API

OpenAI API

Gemini API

Vercel AI SDK

Claude Agent SDK

pgvector

PostgreSQL

Job description

We raised $5.1M to build agents that operate as fully autonomous founders. Our backers include Matt Schlicht (founder of Moltbook), the founders of Crunchbase and Republic, and VCs including Draper, Precursor, and Liquid2. We’re working at the frontier of autonomous systems, laying the groundwork for agents that don’t just chat, but run entire businesses.

The system sources ideas from market signals, validates them with live ad tests, builds and launches software products, and brings in human experts wherever the agents fall short. That loop already runs end to end: from idea, to launched product, to first revenue. The next 12 months are about reliability, throughput, and self-improvement, so that every launch makes the next one better.

We’re hiring a Head of Engineering to own that system: multi-agent orchestration, self-evolving harnesses, the engineering standard, and the team that builds it.

What you'll work on

• Multi-agent systems and orchestration: the agent runtime that takes a product from idea to live ads to launched software, and its reliability, latency, and cost in production
• Self-improving and self-evolving harnesses: the feedback loops, evaluation, and memory that let the system learn from each launch and from human expert input
• Engineering standards: code review, testing, observability, release process, and definition of done
• Building AI-powered backend services on top of LLM APIs (Anthropic, OpenAI, Gemini)

Core responsibilities

• Own the system architecture and make the technical calls; move the team through decisions rather than toward consensus
• Break a large technical problem into the first wins that build momentum, and sequence the roadmap around them
• Ship production code yourself, especially on the agent runtime and orchestration
• Debug and improve AI behavior in production; drive prompt quality, structured outputs, retries, and fallbacks

Required skills and experience

• 4+ years of professional software engineering experience, including leading engineers at an early-stage company (founding engineer, tech lead, head of engineering, or similar)
• Hands-on experience at the bleaching edge, specifically multi-agent systems and orchestration, and self-improving or self-evolving agent harnesses, shipped to production
• Strong backend engineering skills, including designing, building, and maintaining APIs and data flows
• Experience with prompt engineering, model configuration, structured outputs, embeddings, retrieval, and evaluation
• Daily, hands-on use of AI coding agents (Claude Code, Cursor, OpenAI Codex or similar). You plan work, hand it to agents, review what they produce, and know when to verify instead of trust

Nice to have

• TypeScript and Node.js end to end; our stack is TypeScript throughout
• Vercel AI SDK, Claude Agent SDK, or experience building a custom agent runtime
• pgvector on Postgres/Supabase; background job systems such as Inngest or Pub/Sub
• Experience growing or restructuring a distributed team
• Interest in AI evaluation, safety, and observability