Research Intern - Coding Agents & Post-Training Optimization

Microsoft • Redmond, WA

6d • $6,810 - $14,610

About The Position

Microsoft is hiring a Research Intern to improve coding agents through post-training methods, including reinforcement learning, fine-tuning, distillation, compression, and inference optimization. The team trains, evaluates, and ships agentic coding models that power developer tools at scale. These agents plan, edit repositories, run tests, and address failures across multi-step trajectories, making reliability and affordability important goals.

The intern will own a research question end to end: form a hypothesis, build an experiment, evaluate it on realistic agent benchmarks, and advance results toward a research paper or product integration. The role involves working with researchers and engineers across model training, agent infrastructure, and hardware acceleration.

The 12-week internship pairs interns with mentors. Interns collaborate with other interns and researchers, present findings, and participate in the research community. Research Interns are expected to be physically located at their manager’s Microsoft worksite for the duration of the internship. The role is offered year-round, though internships typically begin in the summer.

Requirements

• Currently enrolled full time in a doctoral degree program in Computer Science, Computer Engineering, Electrical Engineering, or a related STEM field.
• Be physically located at the manager’s Microsoft worksite for the duration of the internship.
• Submit at least two reference letters, a cover letter, and relevant work or research samples.
• Research experience in at least one of LLM post-training (reinforcement learning, supervised fine-tuning, or preference optimization), model compression, or efficient inference.
• Proficient Python and PyTorch skills, with hands-on experience training or fine-tuning transformer models.
• Clear communication, self-direction, curiosity, and a bias toward hands-on experimentation.
• Experience collaborating across teams, disciplines, and cultures.

Nice To Haves

• At least one additional quarter or semester of school remaining after the internship.
• First-author publications at top venues (NeurIPS, ICML, ICLR, ACL, EMNLP, MLSys, ICSE, FSE) or equivalent research impact.
• Experience with large-scale RL for LLMs using frameworks such as TRL, verl, OpenRLHF, or DeepSpeed-Chat.
• Experience building or training code-generation models or coding agents, and familiarity with benchmarks such as SWE-bench, LiveCodeBench, Terminal-Bench, or HumanEval.
• Hands-on work with quantization, distillation, pruning, or speculative decoding; familiarity with serving stacks such as vLLM, SGLang, TensorRT-LLM, or ONNX Runtime.
• Familiarity with distributed training (FSDP, DeepSpeed, Megatron) and GPU profiling or kernel work (CUDA, Triton).

Responsibilities

• Own a research question end to end, from forming a hypothesis through building and measuring an experiment.
• Evaluate research results on realistic agent benchmarks.
• Advance research results toward a paper or product integration.
• Collaborate with researchers and engineers across model training, agent infrastructure, and hardware acceleration.
• Work with an assigned mentor and collaborate with other Research Interns and researchers.
• Present findings and contribute to the research community.

What This Job Offers

• Career Level: Intern
• Education Level: Ph.D. or professional degree