Researcher
Bangalore, Karnataka, India
Save
Overview
The PROSE team is advancing the state of the art in program synthesis and shipping these innovations through flagship Microsoft products that impact millions of users worldwide. These products include the world’s most popular spreadsheet software Excel, the industry-leading IDEs Visual Studio and Visual Studio Code, the power platform suite of products including Power BI, PowerApps, PowerAutomate, and other software tools including Azure Data Studio, Azure Data Monitor, Azure Data Factory, PowerShell, and SQL Server Management Studio. The application domains range from developer productivity, business user low-code/no-code scenarios to even programming education. To power these features PROSE has developed a powerful and flexible framework for program synthesis that leverages techniques from both logical reasoning and machine learning (including LLMs) and combines them in unique ways to synthesize programs automatically. This approach has been recognized with frequent papers at top-tier conferences and awards, including best paper and test-of-time awards.
We invite early-career researchers having experience with machine learning or related areas to apply for a full-time position on the PROSE team at Microsoft. The recent success of large language models (LLMs) creates an opportunity to take a qualitative step forward, to extend dramatically the reach of what users can do with Copilots. Our team provides a unique opportunity to work at the intersection of cutting-edge AI research and real-world impact, solving customer problems on a global scale. Associcate researchers collaborate with other researchers to design and carry out the research underlying state-of-the-art systems, and work together with world-class engineers to implement these systems in production. To learn more about our approach and culture see this award talk on the story behind one of the team’s early innovations: the Flash Fill feature in Excel.
Qualifications
- MSc+ in Computer Science or a relevant field.
- At least 2 publications in top-tier machine learning, programming language, software engineering or relevant venues (e.g. AAAI, NeurIPS, ICML, ICLR, EMNLP, NAACL, ACL, POPL, PLDI, OOPSLA, ICSE, FSE, and comparable).
- Ability to develop an original research agenda, demonstrated by leading at least one publication in a top-tier machine learning, programming language, software engineering or relevant venue.
- At least 2 years of prior research experience in machine learning, programming language, or software engineering.
Preferred Qualifications
- Experience in training transformer-based models (including collecting and curating training data) as part of research projects in an industrial or academic setting.
- Experience working with large language models, including automated prompt tuning, fine-tuning, instruction-tuning and continued pre-training.
- Experience with reinforcement learning and its applications to training LLMs.
- Have interests in applying machine learning to structured and semi-structured data such as tabular data.
- Hold previous industrial research internship positions.
- Experience collaborating across multi-disciplinary or multi-role teams.
- Strong communication skills.
Responsibilities
As an associate researcher, you will have the opportunity to:
- Design and carry out research jointly with other researchers, engineers, and PMs.
- Work with engineers and PMs to deploy your research to customers.
- Collaborate with other researchers and longer-term research fellows.
- Publish in top-tier academic venues.
