Research Intern - AI Frontiers - Reasoning & Agentic Models

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

Research Intern position focused on advancing agentic model capabilities, including reasoning, tool use, and interaction across text and visual environments. The role involves developing novel training algorithms, exploring synthetic environments, multi-agent training, and scaling laws, with a focus on foundational models and modern architectures. The internship offers a chance to publish findings and potentially ship AI technologies in products.

What you'd actually do

  1. Advance the state of the art in agentic model capabilities -- creating models and agents that can reliably perform tasks across digital systems on behalf of humans, combining automation, reasoning, and interaction capabilities to execute workflows end-to-end, leveraging both text-based environments (CLI tools, APIs, scripts, MCPs) and visual environments (GUI applications).
  2. Conduct cutting-edge research in artificial intelligence (AI) and publish findings in top-tier venue such as NeurIPS, ICLR, ICML, and others.
  3. Release models and libraries in the open-source (e.g. Phi-4, Phi-4-reaoning, AgentInstruct, AutoGen, MagenticOne) while also working within Microsoft's ecosystem to ship our AI technologies in multiple products.
  4. Develop original research and perform hands-on research in a collaborative and dynamic environment.

Skills

Required

  • deep learning
  • large language model training

Nice to have

  • language model pre-training
  • post-training
  • reinforcement learning
  • original research
  • hands-on research
  • collaborative environment
  • dynamic environment

What the JD emphasized

  • demonstrated ability for technical work
  • proven record of influential publications on Artificial Intelligence
  • Proven publication record in top-tier conferences

Other signals

  • advancing the state of the art in agentic model capabilities
  • creating models and agents that can reliably perform tasks across digital systems on behalf of humans
  • combining automation, reasoning, and interaction capabilities to execute workflows end-to-end
  • leveraging both text-based environments (CLI tools, APIs, scripts, MCPs) and visual environments (GUI applications)