This list contains only the countries for which job offers have been published in the selected language (e.g., in the French version, only job offers written in French are displayed, and in the English version, only those in English).
The AI Frontiers lab at Microsoft Research is charted with ambitious research goals for advancing Artificial Intelligence (AI) capabilities in several key areas including modeling, algorithms, reasoning and agentic AI. Our lab offers a vibrant environment for cutting-edge multidisciplinary research, including an open publication policy and close links to top academic institutions around the world. This Principal Research Engineer position is a unique opportunity to contribute towards tackling some of the hardest and most rewarding challenges in AI. You will help develop novel ideas in bleeding edge reinforcement learning research as well as help evolve our pre-training, mid-training, and post-training codebases that gave birth to famous models such as Phi establishing many new records. You will collaborate with researchers and engineers across many disciplines to help advance the state of the art in reasoning and agentic AI.
Job Responsibility:
Design, develop, execute, and implement technology research projects in collaboration with other researchers, engineers, and product groups
Be a part of research breakthroughs in the field and play a crucial role in developing, improving, and exploring the capabilities of Large Language Models (LLMs), reasoning and agentic AI
Embody our culture and values
Requirements:
Bachelors in Computer Science or relevant field AND 6+ years related experience
OR Master's Degree in Computer Science or related field AND 4+ years related experience
OR Doctorate in Computer Science or related field AND 3+ years related experience
OR equivalent experience
1+ year(s) experience developing with Python and Pytorch/JAX
Familiarity with architecture and optimizations for large language models
Hands-on work in debugging and profiling Pytorch distributed code
Basic understanding of working of CUDA kernels
Familiarity with pre-training, mid-training and/or post-training pipelines for language and/or multimodal models
Foundational understanding of reinforcement learning and key challenges in the field
Nice to have:
Experience with verl, Ray, Megatron and/or vLLM is a significant plus
Any experience in building scalable services can be highly complementary