Principal Researcher

Microsoft Corporation•Redmond, WA

34d

About The Position

The Microsoft Applied Sciences Group incubates disruptive technologies for Microsoft's next-gen hardware products and is working on several exciting projects that will shape how computers and other devices perceive the user and the user's environment. Operating as a startup within the company, this team works closely with several research and product teams to bring compelling new experiences to the market. A lot of these experiences are powered by machine learning - and as part of this team, you will have the unique opportunity to work on almost every aspect of a machine learning project - from ideation to shipping. You will have the opportunity to turn ideas into reality with real-world impact at Microsoft scale. As Principal Researcher you will have the opportunity to design and lead the implementation of state-of-the-art language and multimodal models to transform the Windows experience. You will have the opportunity to collaborate with people at the forefront of their field from across the organization and see your ideas turn ideas into reality with real-world impact at Microsoft scale. This is a hands-on role. Microsoft's mission is to empower every person and every organization on the planet to achieve more. As employees we come together with a growth mindset, innovate to empower others, and collaborate to realize our shared goals. Each day we build on our values of respect, integrity, and accountability to create a culture of inclusion where everyone can thrive at work and beyond.

Requirements

Doctorate in relevant field AND 3+ years related research experience OR equivalent experience.
Experience conducting research as part of a research program in academic or industry settings.
Ability to meet Microsoft, customer and/or government security screening requirements are required for this role. These requirements include but are not limited to the following specialized security screenings: Microsoft Cloud Background Check: This position will be required to pass the Microsoft Cloud background check upon hire/transfer and every two years thereafter.

Nice To Haves

3+ years' experience training and/or finetuning large language models, model compression, distillation to small language models.
3+ years' experience with Python and deep learning libraries
Experience in large language/multimodal model training, finetuning and distillation.
Demonstrated record of publications in top-tier conferences or journals (ICLR, ACL, EMLP, ICML, CVPR, ICCV, ECCV, NeurIPS, TPAMI, etc.).
Experience in creating reusable software libraries and packages.
Experience using dataset curation, data generation using prompting state of art LLMs and/or model alignment.
Experience with reinforcement learning libraries or formal methods and verification is a plus.
Demonstrated ability and passion for incubating new ideas, solving problems, and building working systems.

Responsibilities

Research focusing on foundation models - Designing and training language and multimodal models (and finetunes) with supervised finetuning, distillation, reinforcement learning, low-rank adaptation.
Research focusing on environment interaction, policy specification and verification and human intent alignment.
Implementation: Train, distill, and finetune language and vision models in PyTorch, DeepSpeed, AzureML stack.
Machine learning engineering - Build pipelines to test designs, algorithms and models.
Data science - Research and develop synthetic data generation strategies.
Proactively follow state of the art research and share latest work, write papers, attend conferences and share knowledge in the wider team.

Stand Out From the Crowd

Upload your resume and get instant feedback on how well it matches this job.

Upload and Match Resume