Founding Engineer

Why This Role We're a YC + SoftBank-backed team that has launched a new venture within HR Tech — and revenue has doubled every month since launch. We're building AI agents that operate autonomously in the real world, including over voice. This isn't a wrapper on top of an LLM The core product relies on multi-step agentic workflows, real-time voice AI, and intelligent matching systems that get smarter with every interaction. We're hiring the engineer who will own these systems end-to-end — architecture, shipping, and iteration — and who will help us scale from rapid traction to category leader. What You'll Own • Design and ship agentic AI pipelines — multi-step, tool-using, self-correcting workflows that operate reliably at scale without human intervention • Build and iterate on real-time voice AI systems: low-latency speech pipelines, turn-taking logic, interruption handling, and voice-to-action flows • Own intelligent matching and discovery systems — the core AI that powers how we connect the right people at the right time • Make high-leverage architectural decisions: when to use RAG vs. fine-tuning, how to structure memory across agent sessions, how to instrument and debug non-deterministic systems • Be a genuine thought partner to the founders on product strategy — shape the roadmap, not just execute it • Set the engineering bar and build the culture as the team scales About you Engineering foundation • 7+ years of experience, including 5+ years shipping Node.js and React in production • Startup-tested: you've worked at a Seed or Series A company, made hard tradeoffs under pressure, and know what it takes to move fast with limited resources • Owned infrastructure end-to-end — CI/CD pipelines in GitHub Actions, Docker deployments to container orchestration, static assets to S3, and AWS services including Lambda, ECS, CloudWatch, and RDS AI systems builder • Has shipped agentic or voice AI products in production — you have real scars from debugging non-deterministic, latency-sensitive systems at scale • Deeply familiar with LLM failure modes: hallucination, context limits, tool-calling reliability — and knows how to build guardrails that actually hold • Fluent in AI tooling across the full development lifecycle, from architecture and planning through to scaling; strong prompt engineer who knows how to get reliable output from models How you work • Product-minded: you think in user outcomes, not tickets, and you push back when the spec doesn't make sense • Highly autonomous — you identify what matters, prioritize ruthlessly, and execute without hand-holding • Wired for speed without sacrificing quality: you ship fast and ship things users love, and you know the difference between acceptable debt and the kind that kills momentum • Strong communicator — equally comfortable in a technical architecture discussion and a conversation with a non-technical client or stakeholder Nice to Have • Experience with real-time audio streaming, voice AI, or telephony integrations (Twilio, Vapi, etc.) • Prior work on RAG pipelines, vector databases. or embedding-based retrieval systems • Bay Area based — not required, but a plus The hard problems here are real: keeping agents on-task, making voice feel natural, building systems that work when the model surprises you. If that's the kind of engineering you want to spend your time on — and you want founding-level ownership doing it — let's talk.