Principal Software Engineer - AI Training Systems (CoreAI)

Microsoft•Redmond, WA
•$142,800 - $331,200

About The Position

Microsoft Foundry helps developers and organizations build, deploy, and operate AI solutions at scale. This role is part of an engineering team building infrastructure that makes model training and inference faster, more reliable, and more efficient for customers. The engineer will own distributed-training performance across pre-training, fine-tuning, and reinforcement learning, improving agent task success through post-training while reducing GPU-hours and training costs.

Requirements

  • Bachelor's degree in Computer Science or a related technical field and 6+ years of technical engineering experience coding in languages including C, C++, C#, Java, JavaScript, or Python, or equivalent experience.
  • Ability to meet Microsoft, customer, and/or government security screening requirements, including passing the Microsoft Cloud Background Check upon hire or transfer and every two years thereafter.

Nice To Haves

  • 8+ years of technical engineering experience coding in languages including C, C++, C#, Java, JavaScript, or Python, or equivalent experience.
  • Production distributed-training delivery and proven compute, memory, communication, and convergence fundamentals.
  • Proven software engineering fundamentals, debugging skills, and clear cross-team communication.
  • Experience building training services used by customers or research teams.
  • Experience with Kubernetes, Ray, and training toolchains such as PyTorch, JAX, Slime, verl, or TRL.

Responsibilities

  • Apply FSDP/ZeRO, tensor, pipeline, and expert parallelism, mixed precision, recomputation, and communication overlap where beneficial.
  • Connect asynchronous agent rollouts and feedback to training, balance resources, bound staleness, and improve recovery.
  • Mentor engineers and validate multi-node scaling against quality targets.
  • Build secure, tenant-isolated systems with automated testing, observability, reliable recovery, and safe releases.
  • Participate in operational support and improve services using production feedback.
  • Preserve model quality and numerical correctness.
  • Use reproducible benchmarks to improve throughput, latency, and cost.
  • Pursue near-linear throughput scaling where workloads permit across agreed GPU and node configurations.
  • Validate convergence and held-out model and agent quality, including multi-turn task completion, tool-use correctness, and reward-exploitation checks for reinforcement learning.

Benefits

  • Certain roles may be eligible for benefits and other compensation; additional benefits and pay information is available through the Microsoft careers pay information page.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service