Principal Applied Scientist

Microsoft Microsoft · Big Tech · Redmond, WA +1 · Applied Sciences

The Principal Applied Scientist will work on the core systems for Microsoft Copilot for enterprise, focusing on advancing orchestrator reasoning, training next-generation models, and shipping model-driven experiences. Responsibilities include evolving the orchestrator for multi-modality and agent-based systems, advancing model training (post-training, data curation, frontier training), and building/delivering model-driven features from concept to production.

What you'd actually do

  1. Evolve our orchestrator to reason more effectively, respond faster, and support new capabilities in multi-modality, agent-based systems, and beyond
  2. Advance the state of the art in model training—including post-training methods such as fine-tuning and reinforcement learning, high-quality data curation, and experimentation with frontier model training
  3. Build and deliver model-driven features from concept to production—spanning model development, evaluation, metrics, and A/B testing
  4. Conduct research and development to push the boundaries of model training, evaluation, and quality assessment for AI solutions.
  5. Post-train LLMs for enterprise scenarios M365 Copilot and on tenant data to enable task-specific agents and solutions within the enterprise ecosystem.

Skills

Required

  • Bachelor's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 6+ years related experience
  • Master's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 4+ years related experience
  • Doctorate in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 3+ years related experience
  • equivalent experience

Nice to have

  • Master's Degree in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 9+ years related experience
  • Doctorate in Statistics, Econometrics, Computer Science, Electrical or Computer Engineering, or related field AND 6+ years related experience
  • equivalent experience
  • 5+ years experience creating publications (e.g., patents, libraries, peer-reviewed academic papers)
  • 2+ years experience presenting at conferences or other events in the outside research/industry community as an invited speaker
  • 5+ years experience conducting research as part of a research program (in academic or industry settings)
  • 3+ years experience developing and deploying live production systems, as part of a product team
  • 3+ years experience developing and deploying products or systems at multiple points in the product cycle from ideation to shipping
  • 4+ years of experience with end-to-end in a challenging technical problem domain (plan, design, execution, continuous release, and service operation)
  • 4+ years experience with shipping internet scale, low latency and high throughput systems AI products
  • 4+ years experience with customer/End-result/Metrics driven in design and development
  • Publications at top conferences like ACL (Association for Computational Linguistics), EMNLP (Empirical Methods in Natural Language Processing), SIGKDD (Special Interest Group on Knowledge Discovery and Data Mining), AAAI (Association for the Advancement of Artificial Intelligence), WSDM (Web Search and Data Mining), COLING (International Conference on Computational Linguistics), WWW (World Wide Web Conference), NIPS (Neural Information Processing Systems), ICASSP (International Conference on Acoustics, Speech, and Signal Processing), etc.

What the JD emphasized

  • pushing the frontier of large language models
  • core systems
  • model training
  • post-training methods
  • frontier model training
  • model-driven experiences
  • model development
  • model performance
  • quality assessment
  • Post-train LLMs
  • task-specific agents
  • multimodal models
  • model optimization techniques
  • model performance
  • quality
  • impact
  • shipping internet scale, low latency and high throughput systems AI products

Other signals

  • core systems that power Microsoft 365 Copilot Chat
  • advancing orchestrator reasoning
  • training next-generation models
  • shipping impactful, model-driven experiences