About the role
Requirements:
- 3+ years professional programming (JD) in web application + backend environments practical target 5-8 years.
- Expert-level TypeScript/JavaScript or Python.
- Strong modern web framework: React, Next.js, or Rails + a backend service framework: FastAPI or Node.js .
- True full-stack: frontend + backend + system architecture.
- Must be confident on both ends; they explicitly refused to define a front-end/back-end split (we don't want 20%/30%, people).
- Even 1 year in FE is fine but should be confident.
- Has shipped a B2B or B2C product using a foundational or open-weight model; this is the hard gate.
- Hands-on building AI and agent-powered features using frontier models (OpenAI, Anthropic, Google), prompting, context, tools, model reasoning.
- AI evals + tracing: Ashish called this out separately as what we're looking at in the majority of our AI hires: evaluating AI agents, making quality measurable and reproducible.
- Understands strengths/limitations of frontier vs open-weight models.
- Daily use of coding agent harnesses Claude Code, Codex, Opencode, Pi as part of their normal SDLC.
- Production engineering hygiene: observable, maintainable, scalable systems; continuous delivery via feature flagging and trunk-based development.
Strong plus:
- Agent-based development (plus, not a hard requirement).
- Using AI models to generate content consumed by end users.