Sr. Machine Learning Engineer

Intel Intel · Semiconductors · California, Santa Clara, United States +3

This role focuses on agent harness research and model fine-tuning, sitting at the intersection of research and engineering. Responsibilities include designing and implementing algorithms for agent harness and post-training pipelines, developing RL environments and reward models, and conducting training runs to improve model capabilities for agentic applications. The role also involves building evaluation benchmarks and metrics, and debugging/optimizing training runs.

What you'd actually do

  1. Build evaluation benchmarks and metrics
  2. Build and iterate on agent harness, including context engineering, agent memory, tools, skills.
  3. Build, maintain, and iterate on the post-training pipeline: Develop robust, reproducible training workflows from data ingestion and preprocessing through model checkpointing and deployment
  4. Design RL environments and reward functions — Develop environments, reward signals, and verifiable reward frameworks for training models on reasoning-intensive tasks.
  5. Debug and optimize training runs — Profile training jobs, resolve bottlenecks, improve GPU utilization, and address numerical instability at multi-GPU scale

Skills

Required

  • BS in CS, EE, Math or related STEM field
  • Python
  • LLM architectures, optimization and model training dynamics
  • Designing and building evaluation frameworks and benchmarks

Nice to have

  • Masters or PhD degrees
  • Implementing and scaling the full post-training pipeline for language models including supervised fine tuning and reinforcement learning
  • Owning and driving a research agenda independently
  • Ambiguity tolerance
  • Debug-first mindset
  • Research-engineering balance
  • Collaborative work style
  • Clear technical communication

What the JD emphasized

  • 8+ years software development background
  • 4+ years of hands-on experience in machine learning engineering, data science or ML research.
  • Experiences designing and building evaluation frameworks and benchmarks that accurately measure model capability improvements and alignment quality
  • Proficient in LLM architectures, optimization and model training dynamics.
  • Hands-on experience implementing and scaling the full post-training pipeline for language models including supervised fine tuning and reinforcement learning.

Other signals

  • agent harness research
  • model fine tuning
  • post-training pipeline
  • RL environments and reward models
  • evaluation benchmarks