Developer Experience Engineer

Black Forest Labs Black Forest Labs · Multimodal · Freiburg, San Francisco · Product

This role focuses on enhancing the developer experience for Black Forest Labs' generative AI platform, FLUX. The engineer will own the end-to-end developer experience, ensuring it's seamless for both human developers and AI agents. This involves working on API contracts, SDKs, dashboards, and agent-specific features to make BFL the preferred platform for visual AI development.

What you'd actually do

  1. Own DevExp as a product surface — wherever the bottleneck is, that's where you ship next. One sprint that might be the API contract, the next it's the dashboard onboarding, the next it's an MCP server, the next it's a Python SDK abstraction
  2. Partner with platform engineering on API design before it ships, not after — bring the developer-and-agent perspective into the room when the contract is being decided
  3. Work with product engineering on the developer-facing surfaces of the dashboard — playground, API keys, usage, request inspectors, webhook debugging — bringing the DevExp lens to features the rest of the team is already shipping
  4. Make BFL the best visual AI platform for agents to use. Own the MCP server, skills, agent onboarding paths, and the machine-readable surfaces of the platform. Design the reference patterns and SDK primitives developers reach for when building agentic workflows on FLUX. Stress-test every API and SDK change against "would Claude / GPT / an autonomous agent get this right on the first try?" — and treat agents as a first-class user across the whole product (errors an LLM can recover from, capabilities discoverable without a human reading docs, audit trails that make sense when most of the traffic on an account isn't a person).
  5. Own the portfolio of entry points developers and agents reach for — SDKs (Python, TypeScript, more on demand), MCP server, skills, OpenAPI / typed surfaces — and treat them as a portfolio, not a hierarchy. SDKs add composable workflows and sane defaults for async / polling / retries; MCP and skills make BFL trivially usable from any agent loop. DevRel and the Creative Technologist team partner on examples, content, and showcase work on top of these surfaces — you own the surfaces themselves.

Skills

Required

  • 4+ years building software, with meaningful time on developer-facing products — SDKs, APIs, dashboards, consoles, platforms developers actually use
  • Full-stack comfortable. TypeScript / React / Next on the front, Python (and ideally Go or Rust) on the back.
  • Deep instincts on API design — you have opinions about error formats, pagination, async job shapes, breaking changes, and you've lived with the consequences of bad ones
  • Fluent in the 2026 agent stack — MCP, tool use, the major agent runtimes — and you know what breaks when an LLM is the one reading your error message or browsing your API
  • You've shipped developer-facing product that real external developers depend on. Not internal tooling.
  • You ship. This is product engineering, not product management with an IDE open.

Nice to have

  • Background at a developer platform where DevExp was a top-level metric (Stripe, Vercel, Cloudflare, Modal, Replicate, Resend, Linear, Anthropic platform, etc.)
  • You've built or meaningfully contributed to an MCP server real users / agents rely on
  • Experience with media-generation or other high-latency, async, job-shaped APIs — and the UX puzzles that come with them (polling, webhooks, partial results, cost surprises)
  • A public surface — OSS, a writeup, a dashboard you're proud of — that lets us see your taste before the interview

What the JD emphasized

  • developer-facing products
  • external developers depend on
  • agent stack
  • developer experience

Other signals

  • developer experience
  • platform engineering
  • agentic workflows
  • API design
  • SDKs