Microsoft Principal Research Engineer – Generative AI Frontiers 

Taiwan, Taoyuan City 

508086778

with ambitious research goalsrtificial Intelligencein several key areasincluding modeling, algorithms,and agentic.lab offers a vibrant environment formultidisciplinary research, including an open publication policy

some of the hardestand most rewardingin AI.ou will help developnovel ideas inreinforcement learningas well ashelp evolve, mid-training,andthat gave birth to famous modelssuch as Phinew records. You will collaborate with researchers andmany disciplinesadvance thestate of the art inreasoning and agentic AI.

Our lab providesopportunities for experimentationand access toarray ofreal-world problems and dataalong withto ship your researchto overexpand thepareto frontierof AI capabilities, efficiency, and safety through innovations in foundation models and learning agent platforms.ur projects include Phi, Orca,, Dion,Belief State Transformersamong many others.

Our ongoing research areas encompass but are not limited to:

• Pre-training: language models, action models and multimodal models
• Mid-training: long context extension
• Post-training: e.g., Instruction tuning and reinforcement learning from feedback
• Reasoning: Enabling LLMs toscale inferencetime compute via reinforcement learning
• Action Models: Trailing models capable oftaking actionin the digital world (e.g.computer use and web agents)
• Orchestration and multi-agent systems: automated orchestration between multiple agents incorporatingcollaboration, humanfeedbackand oversight

If yourskills and

Required Qualifications

• Bachelors in Computer Scienceor relevant field AND6+ years relatedexperience
  • ORMaster's Degree inComputer Scienceor related field AND4+ years related experience
  • OR Doctorate inComputer Science or related fieldAND 3+ years related experience
  • OR equivalent experience.
• 1+ year(s)experiencedevelopingwithPython andPytorch/JAX

Preferred Qualifications

• Familiarity witharchitecture and optimizations for large language models.
• Hands-on workin debugging and profilingPytorchdistributed code.
• Basic understanding ofworkingofCUDAkernels.
• Familiarity withpre-training, mid-training and/or post-training pipelines forlanguageand/or multimodal models.
• Foundational understanding ofreinforcement learning andkeychallenges in the field.
• Experience withverl, Ray, Megatronand/orvLLMis a significant plus.
• Any experience in building scalable servicescanbe highlyc

Certain roles may be eligible for benefits and other compensation. Find additional benefits and pay information here:

• As aPrincipal Research Engineerin AI Frontiers, you will design, develop, execute, and implement technology research projects in collaboration with other researchers, engineers, and product groups.
• As a member of aworld-class research organization, you will be a part of research breakthroughs in the fieldandplay a crucial role in developing, improving, and exploring the capabilities of Large Language Models (LLMs), reasoning and agentic AI.

• Embody our and .