Research Platform Engineer

What You Will Do

• Design and build training infrastructure, data infrastructure, and data processing and sourcing pipelines.
• Productionize models for serving and own parts of the inference stack.
• Build internal tools and services that increase engineering and research velocity.
• Debug hard problems across training, inference, and performance — including distributed systems issues under real research workloads.
• Optimize throughput, latency, GPU utilization, and system-level scaling.
• Improve research iteration speed and developer experience — cut debugging time, raise reliability, and make it faster for researchers to ship experiments.
• Raise the engineering bar across research and platform code alike.

Key Qualifications

• 5+ years of experience building and shipping production systems, with demonstrated ownership of infrastructure used by other engineers or researchers.
• Strong depth in at least one of: ML infrastructure, distributed training or inference systems, data systems, or research tooling.
• Strong distributed systems foundations — concurrency, consistency tradeoffs, replication, failure modes, and scaling behavior under real workloads.
• Strong performance optimization skills across at least one of: training throughput, inference latency, GPU utilization, or system-level scaling.
• Strong proficiency in Python, with the ability to work in C++, CUDA, Rust, or Go as the work demands.
• Experience working directly with ML researchers or research engineers, including productionizing research code.
• A product engineer's instincts for iteration speed and developer experience — applied to the systems researchers use every day.
• Strong judgment about what to build and what to leave alone, particularly when research requirements are ambiguous or shifting.
• High-ownership mindset; you measure yourself by outcomes shipped, not by tickets closed.

Preferred Qualifications

• Experience at an AI lab or ML-native company, working on systems used directly by researchers.
• Experience with large-scale training or inference systems — GPU orchestration, distributed training, or high-throughput inference.
• Experience with low-level performance optimization — profiling, kernel-level tuning, memory and bandwidth optimization, distributed communication primitives.
• Experience building developer experience tooling for research — notebooks, experiment tracking, reproducibility infrastructure.
• Experience in early-stage or high-growth environments where scope and priorities shift frequently.