The Copilot Platform AML Team is driving the next generation of intelligent assistant infrastructure, powering Microsoft Copilot experiences across the enterprise. Our mission is to build foundational language models that make Copilot more helpful, responsive, and accessible to millions of users worldwide.
We are looking for Applied Scientists to pioneer innovations in scalable training and inference optimization for both Small and Large Language Models (SLMs/LLMs). In this role, you will directly shape the core platform capabilities of Copilot, influencing how organizations interact with AI-driven assistants every day.
Our work spans the entire model lifecycle—from supervised fine-tuning to advanced post-training techniques such as instruction tuning, reinforcement learning, and alignment. We also push the boundaries of model efficiency with cutting-edge compression strategies, including GPTQ, AWQ, and pruning, to deliver faster, more cost-effective inference at scale.
If you’re passionate about creating intelligent assistant systems that combine deep model expertise with world-class engineering, and want to shape the future of enterprise AI, we’d love to have you on our team.

Basic Qualifications:
Master’s degree or above (or equivalent experience) in Computer Science, Engineering, Mathematics, Physics, or a related field.
Strong programming skills with hands-on experience in managing large-scale data and machine learning pipelines.
Deep understanding of open-source ML frameworks such as PyTorch, vLLM, and TensorRT-LLM (TRT-LLM).
Solid knowledge of model optimization techniques, including quantization, pruning, and efficient inference.

Preferred Qualifications:
1+ years of experience optimizing LLM inference using frameworks like vLLM or TRT-LLM.
Practical experience in model compression and deployment within production systems.
Experience designing agentic AI systems, such as multi-agent orchestration, tool usage, planning, and reasoning.

Model Optimization & Deployment:
Design and implement efficient workflows for training, distillation, and fine-tuning Small and Large Language Models (SLMs), leveraging techniques such as LoRA, QLoRA, and instruction tuning. model compression strategies—including quantization (e.g., GPTQ, AWQ) and pruning—to reduce inference costs and improve latency.
Optimize LLM inference performance using frameworks like vLLM and TensorRT-LLM (TRT-LLM) to enable scalable, low-latency deployment.
Build robust and scalable inference systems tailored to heterogeneous production environments, with a strong focus on performance, cost-efficiency, and stability.

Evaluation & Data Management：
Develop evaluation datasets and metrics to assess model performance in real-world product scenarios.
Build and maintain end-to-end machine learning pipelines encompassing data preprocessing, training, validation, and deployment.
Cross-functional Collaboration：
Collaborate closely with product managers, engineers, and research scientists to translate business needs into impactful AI solutions, driving real-world adoption and seamless product integration.