AI Fullstack Engineer

AI Fullstack Engineer at Humwork (Uplers Client)

Uplers is seeking an exceptional AI Fullstack Engineer for their client, Humwork, a YC Spring 2026 batch company. Humwork is pioneering a real-time marketplace where AI agents can call human experts when they encounter difficulties, akin to Waymo's remote driver assistance but for AI coding agents. This role offers the unique opportunity to own the full technical spectrum, from AI intelligence to real-time infrastructure and mobile application development.

About Humwork

Humwork is building the real-time marketplace where AI agents call human experts when they get stuck. Think Waymo's remote driver assistance—but for AI coding agents. When Claude Code or Cursor hits a wall, our MCP server connects it to a vetted human expert in seconds. We're in YC's Spring 2026 batch with two founders, moving fast. We're looking for a single exceptional engineer who can own the full picture: from the AI intelligence layer that decides when agents need help, to the real-time infrastructure that connects them with experts, to the mobile app that puts jobs in experts' hands.

What You'll Do

AI & Agent Intelligence

• Autonomous Escalation Systems: Design and implement the core intelligence that detects when coding agents need human help—balancing accuracy with appropriate escalation frequency.
• MCP Server Architecture: Build robust, scalable MCP server integrations across Cursor, Claude Code, Lovable, and other coding agents.
• AI-to-Human Handoff Pipeline: Develop and optimize the real-time, low-latency handoff from AI agent to human expert—prompt engineering, tool descriptions, trigger conditions, context packaging.
• AI Agent Orchestration: Design sophisticated prompting and tool-calling systems that enable natural conversation between AI agents and human experts.
• Experimentation & Benchmarking: Run A/B tests on prompts, benchmark agent performance against baselines, and optimize call frequency and resolution quality.

Platform & Infrastructure

• Real-time Communication Systems: Architect WebSocket infrastructure supporting thousands of concurrent AI-expert sessions—reconnects, message ordering, connection scaling, timeout handling, and message reliability.
• Scalable Backend: Design and build the FastAPI backend end-to-end—expert matching, session management, billing, analytics, and API design.
• Expert Routing Engine: Create algorithms that match coding problems to the most suitable available expert based on skillset, language, timezone, availability, and problem context.
• Database & Data Layer: Own schema design, complex queries, connection pooling, real-time data synchronization, and caching strategies with PostgreSQL and Redis.

Product & User Experience

• Expert Mobile App: Build and ship the expert-facing mobile app (React Native/Expo) with Uber-style push notifications, real-time chat, and availability controls.
• Web Dashboard: Develop the web-facing dashboard (Next.js) for analytics, onboarding, and session management.
• Expert Marketplace Platform: Implement onboarding flows including identity verification, skill assessment, credential verification, and availability scheduling.
• Payments: Integrate Stripe Connect for multi-party payment flows—expert payouts, platform fees, and billing.

You Should Apply If

• You use AI coding agents (Claude Code, Codex, Cursor) daily and are a power user of it.
• You've built with MCP servers, LLM tool-calling, or AI agent frameworks.
• You understand prompt engineering at a deep level—not just writing prompts, but designing system instructions that shape agent behavior.
• Expert-level Python: FastAPI, async programming, and LLM API integrations.
• Strong experience with React/Next.js and building production web applications.
• You've built real-time systems with WebSockets—and you know how to handle reconnects, message ordering, and connection scaling.
• Comfortable with PostgreSQL and ORMs (SQLAlchemy or similar), plus Redis for caching, pub/sub, and queue management.
• You've shipped mobile apps (React Native preferred) or PWAs with push notifications.
• Cloud infrastructure experience: AWS/GCP deployment, containerization, CI/CD.
• You can prototype fast and know when to reach for Opus 4.6 vs. 5.3 Codex.
• You thrive at the intersection of product, research, and engineering—and can own features end-to-end from schema design to shipped UI.

Bonus Points

• You've contributed to or built MCP servers/plugins.
• Experience with autonomous agent frameworks (CrewAI, AutoGen, etc.).
• You've built marketplace or on-demand matching systems (expert routing, driver dispatch, etc.).
• Familiarity with Stripe Connect or multi-party payment flows.
• You've benchmarked AI agent performance or worked on evals.
• Background in human-computer interaction, developer tools, or gig economy platforms.
• Open-source contributions in relevant technologies.

What We Offer

• Equity: Ownership in a Y Combinator company with a clear path to liquidity.
• Ownership: You'll be the engineering team along with the technical founders, involved in every system, decision, and architecture call.
• Impact: Define how AI agents and humans collaborate.
• Growth: Work directly with YC partners and experienced founders at a company building for a $100B+ market.
• Flexibility: Remote-first, core hours overlap with US Pacific Time.

Interview Process

• Interview with founders.
• Technical screening (AI systems + WebSocket architecture + database scaling).
• AI Code review and pair programming session.

Apply With

• GitHub showcasing relevant projects.
• Brief note on why AI-human collaboration excites you.
• Your experience with coding agents (Cursor, Claude Code, etc.).

We're especially interested in candidates who are excited about the intersection of AI and human expertise—and who want to build the infrastructure that makes both better.

Key skills/competency

• Python
• FastAPI
• React Js
• Next Js
• React Native
• Websockets
• PostgreSQL
• Stripe Connect
• AI Coding Agents
• MCP Integration