Why we're hiringWe're doubling down on creation quality and reliability. This is a key role with real ownership: you'll run the lane end-to-end and set the technical bar for agent-driven app generation. We want people who think in systems, not solutions. How do we generate fully autonomous software for recreational use? How do we generate it fast? And how do we do it tastefully? This role lives at the bleeding edge of what current AI can do - if those questions excite you, it's for you.
You will own- Designing and owning our agent harness layer Building long-horizon agentic workflows with maximal autonomy
- Developing robust evaluation and quality loops, including eval harnesses, regression testing, and failure taxonomies
- Designing model strategy: routing, benchmarking, reliability, and cost/latency optimization Creating debuggable agent systems with tracing, metrics, alerts, and observability Researching novel approaches and shipping them to production
We're looking for- You've built agentic systems - tool use, orchestration, retry/repair loops, context management - whether on raw provider APIs or frameworks like LangGraph, PydanticAI, Google ADK, Pi-mono, or Vercel AI SDK.
- You understand the tradeoffs well enough to know when the framework gets in your way. Experience with distributed / real-time systems
- A research-and-ship mindset: you treat open-ended problems as experiments, and you care as much about getting the result into production as discovering it Strong software engineering fundamentals, with depth in Python and/or TypeScript
- A track record of shipping and operating production systems A history of owning hard problems end to end - you can point to systems you built and the outcomes they drove, not just the tech you touched
- Strong product instincts: you can translate user-experience goals into system-design decisions
Bonus points - Took a customer-facing product from 0 to 1
- Built a coding agent, devtool, IDE assistant, or code-generation pipeline (directly relevant - this is what we do)
- Built evaluation frameworks (automated scoring, regression suites, release gating)
Designed MCP-style tool protocols or robust tool interfaces for LLM systems - Experience training and/or fine-tuning LLMs
Compensation & benefits- Competitive compensation: salary + equity
- Full medical / health insurance coverage health insurance
If you are excited about building at the frontier of AI, consumer social, and interactive entertainment, we would love to hear from you.
At Sekai, you will work with an AI-native team, move fast, own meaningful problems, and have the flexibility to explore new ideas from an early stage.
