Principal Applied Scientist

Microsoft Microsoft · Big Tech · Redmond, WA +4 · Applied Sciences

This role focuses on building and improving AI agent platforms, specifically the Agent Performance team within Azure AI Platform. The core responsibility is to integrate scientific advancements, particularly RL techniques like RLHF, into production-ready features for AI agents. The role involves monitoring, evaluating, optimizing, and enabling self-improvement of these agents, bridging the gap between research and product.

What you'd actually do

  1. Help to shape the direction of Agent Foundry Platform and Agent self-improvement with industry-leading product work.
  2. Collaborate with and bridge the gaps between researchers (e.g., across CoreAI, Microsoft Research [MSR] and open-source communities) to translate applied research into differentiated, production-quality features
  3. Bring new technology and approaches, such as Reinforcement Learning from Human Feedback (RLHF), into production by applying long-term research efforts to drive Agent self-improvement
  4. Drive negotiations across teams to ensure cutting edge technology is being applied to products in a practical way that meets key business objectives
  5. Leverage and/or construct data and experimentation to rapidly iterate on and refine product opportunities in a rapidly evolving domain
  6. Proactively provide mentorship and coaching to less experienced and mid-level team members by sharing expertise to build team capabilities and guiding team members in projects, and their careers

Skills

Required

  • Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 6+ years related experience OR Master's Degree AND 4+ years related experience OR Doctorate AND 3+ years related experience OR equivalent experience
  • Ability to meet Microsoft, customer and/or government security screening requirements

Nice to have

  • 4+ years foundation in machine learning and applied AI techniques, including model evaluation and optimization.
  • 3+ years of experience programming in Python.
  • 1+ years of experience in fine-tuning and post-training strategies for models in production environments.
  • 2+ Experience with prompt engineering and optimization for large language models or similar AI systems
  • 3+ years experience presenting at conferences or other events in the outside research/industry community as an invited speaker.
  • 7+ years experience conducting research as part of a research program (in academic or industry settings).
  • 5+ years experience developing and deploying live production systems, as part of a product team.
  • 7+ years experience developing and deploying products or systems at multiple points in the product cycle from ideation to shipping.

What the JD emphasized

  • production-quality features
  • production environments
  • live production systems
  • product cycle from ideation to shipping

Other signals

  • agent platform
  • AI agents
  • evaluate
  • optimize
  • self-improve
  • RLHF
  • production quality features