We launched recruiters.rip, let your AI apply for you. First month free with code onemonthfree

Senior Staff Software Engineer, ML Training Infrastructure

Stack AV is a SoftBank Group subsidiary developing AI-powered autonomous trucking systems.

Pittsburgh, United States
About Stack AV Co.

Stack AV develops autonomous trucking technology using AI, robotics, machine learning, and cloud technologies to improve freight safety, reliability, and efficiency.

View jobs by Stack AV Co.

Skills

Candidate Availability

About the Role

You will drive the design and development of a high-performance multi-tenant AI training platform, balancing hands-on coding with long-term technical direction. You will work across ML Platform, Infrastructure, Autonomy and Safety Evaluation teams, deliver end-to-end ML pipelines, resolve performance bottlenecks, optimize distributed training throughput, and build tooling that helps ML engineers iterate quickly.

Requirements

  • Bachelor's or Master's degree in Computer Science, Engineering, or a related field
  • 8+ years of experience with ML platforms and building ML-based applications
  • Proven track record optimizing training with distributed training paradigms
  • Experience productionizing large-scale data pipelines and data loaders
  • Strong Python skills and experience with Lance, PyTorch, Ray, Spark or equivalent
  • Proven track record building scalable, reliable infrastructure working with MLEs
  • Deep understanding of design tradeoffs and ability to build alignment across cross-functional teams
  • Strong analytical and problem-solving skills
  • Excellent verbal and written communication skills

Responsibilities

  • Design and evolve training platform components including orchestration, scheduling, training abstractions, control plane, observability and performance tuning
  • Deliver end-to-end ML model pipelines across logs processing, feature extraction, dataset schema design and storage, model configuration management, training, and profiling and acceleration
  • Instrument and analyze training infrastructure performance to resolve bottlenecks
  • Optimize training throughput across distributed nodes using topology-aware frameworks and hardware such as InfiniBand, NVLink and NCCL
  • Evangelize system abstractions and tooling that enable MLEs to iterate rapidly
  • Incorporate OSS tools so ML engineers can profile and optimize their workflows
  • Promote engineering excellence within the team