Senior Python Engineer

Role overview This senior role centers on the Python systems that power agent orchestration and data work: multi-agent systems running long-horizon tasks, retrieval pipelines, evaluation harnesses, and the ingestion jobs that feed them. The expectation is to make architectural decisions that other engineers build on, and to defend those decisions in review. The work moves from prototype to production, including the operational concerns that come with it.

Responsibilities - Design and ship multi-agent orchestration, covering task decomposition, tool use, state persistence, and recovery when a run fails halfway through. - Build retrieval and evaluation infrastructure, including eval harnesses that measure whether a change actually improved anything. - Own the LLM integration layer: prompt and context management, token budgets, caching, and fallback behavior under provider failure. - Build ingestion and transformation pipelines that process large volumes without silent data loss, designing for idempotency and replay. - Choose between Postgres, vector stores, and object storage based on evidence rather than habit. - Set standards for typing, testing, and packaging across the Python codebase, and mentor engineers through review and design discussion.

Requirements - Six or more years writing production Python, with systems you designed and owned. - Strong async Python, type hints, and a working understanding of runtime performance characteristics. - Production experience with LLM systems, including retrieval, agent frameworks, or inference serving. - Comfort with cloud infrastructure on GCP or AWS, including deploying and operating what you build.

Nice to have - Experience with vector databases and embedding pipelines at scale. - Model serving or fine-tuning experience, including local inference. - Background in distributed systems or large-scale concurrent processing.