Senior Full Stack Engineer (AI)

Build the AI agents that run the back office for food supply chains.

Burnt builds AI that runs order entry, procurement, pricing, and customer communication for food distributors: the manual work that eats 4 to 5 hours of a sales rep's day and most of an ops team's week. Our customers are mid-market and large distributors that move food from manufacturers to restaurants and stores.

We're a small team growing fast, with large national accounts going live this year. The role is an opportunity to have a significant impact on the product, helping to shape the future of the company as we scale. This is San Francisco, in person.

The role
You'll build the AI systems and the product surface our customers run their business on. Food distribution is messy: every distributor runs a different ERP, orders arrive by email, phone, EDI, and PDF, and the data is never clean. You'll design and ship the agents that turn that into structured, reliable action, plus the features distributors use every day. You'll own systems end to end, from the React frontend to the services behind it, to the model in the middle, and you'll be on call when they run in production.

We want engineers who were writing production code before LLMs existed and who now use AI to move faster, not to cover gaps. We are looking for people who are comfortable ownining the production process end to end.

Tech stack (mandatory)

• Node.js, TypeScript, NestJS, React, Terraform
• AWS (any combination of Lambda, ECS, RDS, S3, and similar)
• LLM / agent frameworks (LangChain, LlamaIndex, or equivalent)
• Observability tooling (Datadog, OpenTelemetry, CloudWatch, or similar)

What you've done

• Built and deployed production-grade AI agents, not prototypes
• Architected and shipped full-stack features live in production
• Handled large-scale datasets at the application layer
• Designed and implemented eval frameworks for AI systems
• Built self-learning or feedback-loop systems
• Treated observability as first-class: logging, tracing, and alerting baked in from day one
• Owned real production incidents and system failures start to finish
• Comfortable using AI coding tools (Cursor, Claude Code, Copilot) to multiply your output
• Wrote production code before the LLM-assisted era and know what good looks like without it

What you'll need

• NestJS, TypeScript, and React in production
• At least one AI agent shipped and running in production
• AWS applications deployed with real traffic
• Personally debugged and resolved production failures
• Can articulate system design decisions you personally architected
• Pre-LLM coding experience, with foundational engineering instincts rather than vibe coding

Logistics
In person in San Francisco.