Senior AI Platform Engineer

We are looking for a Senior AI Platform Engineer to build and scale the infrastructure behind production AI agents. Our AI agents conduct live conversations — including training role-plays, outbound calls, and interviews — and turn conversations into scores, summaries, and coaching insights. What you'll work on AI Agents: Build and run production voice/conversational agents, including state, interruptions, latency, configuration, versioning, and testing. Durable Workflows: Design event-driven pipelines using queues, webhooks, retries, and idempotent processing. LLM Pipelines: Build multi-step analysis pipelines using frontier and open-source models, balancing quality, latency, and cost. Evals & Observability: Build tracing, evaluations, and quality gates for agents, prompts, and models. Architecture & Data: Own technical decisions and design secure multi-tenant systems using PostgreSQL. Engineering: Review code, write design notes, improve architecture, and help other engineers grow. What we're looking for Production experience building and operating AI agents with real users . Experience with voice agent platforms or open-source agent frameworks . Strong understanding of latency, interruptions, state, and production failure modes. Production experience with durable workflows and event-driven systems . Hands-on experience with retries, queues, webhooks, and idempotency. Experience shipping LLM-powered features with frontier or open-source models. Experience with LLM observability, tracing, and evals . Strong PostgreSQL skills, including schema design and migrations. Strong TypeScript and comfortable with Python . Strong written English and solid technical communication. Nice to have Speech/transcript pipelines at scale. Speaker diarization and long-form audio processing. Open-source voice agent frameworks. SIP, telephony, phone carriers, or WebRTC. Building eval sets that gate prompt/model changes. Next.js and Supabase. Open-source contributions to AI, voice, or LLM tooling. How we work AI-assisted coding is expected — we care about your ability to verify and judge AI-generated code . Written-first: design notes, clear PRs, and async communication. PR + CI for every change. Staging before production. Regular code reviews and pairing. Small team, high ownership, direct impact on architecture and product. Why join You'll own a meaningful part of the production AI stack , from live voice interaction to durable processing, LLM analysis, evaluation, and the final product experience. If you match most of the requirements but not all of them, apply anyway .