Principal Engineer, Model Development Platform

Responsibilities

• System architecture & reliability - Design and evolve the platform's overall architecture for reliability, observability, and scalability. Set performance, latency, and availability targets, and drive the engineering standards to meet them.
• Cross-domain technical leadership - Unify the platform across disciplines, from front-end UIs and distributed training to Spark data pipelines and optimization-based experiment scheduling, ensuring systems interoperate cleanly.
• Hands-on problem solving - Dive into the hardest challenges across subteams, lead architectural reviews, and propose pragmatic solutions that balance innovation with operational simplicity.
• Experimentation & scheduling systems - Build systems that optimize how models are tested in simulation and on-road, using techniques like linear programming and heuristic optimization to balance hardware, safety, and research priorities while improving throughput and turnaround.
• Data & compute infrastructure - Architect pipelines that ingest, transform, and enrich petabytes of fleet sensor data, and drive efficient compute use across GPU, CPU, cloud, and edge for both prototyping and large-scale training.
• Strategic collaboration - Partner with Product, Research, and Operations to align architecture with user needs and co-own the platform's long-term roadmap.

Essential

• Technical Leadership at Scale – 10+ years of experience designing and building large-scale distributed systems, ML/AI infrastructure, full stack web application, or developer platforms, including at least 3 years as a staff or principal-level engineer.
• Architectural Depth & Breadth – Proven ability to design systems spanning web platforms, ML pipelines, and large-scale compute orchestration (e.g., Spark, Ray,
  Kubernetes
  , Airflow, MLflow).
• Reliability and performance – Experience driving platform reliability improvements, defining SLAs/SLOs, and building self-healing and observable systems that operate at “four nines” availability or better.
• Hands-On Systems Design