Infrastructure Engineer

You build the platform the whole operation runs on, and you're the rare role that spans both halves of it: the AI system that keeps thousands of agents running experiments, and the trading systems that take the strategies they find to live markets. One week that's schedulers and reproducibility for research, the next it's market data feeds and execution. At our scale, most of it is built in-house. WHAT YOU'D DO - Run the scheduler and orchestration that keep thousands of concurrent agent jobs alive, with checkpointing and recovery when they fail - Build the data layer that feeds them: market data feed handlers and research pipelines, from ingestion to clean, versioned datasets - Make every result reproducible and traceable, from a config hash back to the run that produced it - Take a validated strategy to market: exchange connectivity, order execution, and the monitoring to run it safely YOU MIGHT BE A GOOD FIT IF - You have deep experience building and operating distributed systems in production (what you've shipped matters more than titles or degrees) - You're fluent in Python and comfortable across Go, Kubernetes, and distributed compute or streaming frameworks like Ray, Spark, or Kafka - You treat reliability as a discipline: monitoring, incident response, and not failing the same way twice - You like range, and would rather work across research infrastructure and trading systems than specialize in one corner BONUS - Low-latency or trading systems: market data feed handlers, exchange connectivity, or order execution - Running large ML or research workloads on GPUs, with scheduling, checkpointing, and fault tolerance DETAILS - Stack: Python, Go, Kubernetes, streaming data, low-latency systems - Location: San Francisco (in-person) - Comp: $175K to $250K base, plus 0.25% to 0.75% equity