Research Intern - Microsoft Teams CMD Labs

Microsoft • Redmond, WA

276d • $5,460 - $10,680 • Remote

About The Position

Research Internships at Microsoft provide a dynamic environment for research careers with a network of world-class research labs led by globally-recognized scientists and engineers, who pursue innovation in a range of scientific and technical disciplines to help solve complex challenges in diverse fields, including computing, healthcare, economics, and the environment. Microsoft Teams is the hub for teamwork that integrates all the people, content, and tools your team needs to be more engaged and effective. It is core to Microsoft's modern work, modern life & modern education value prop. We are reinventing the way people communicate and work together across the globe. We are looking to hire a PhD (or published MSc) candidate for a 12-week Research Internship to join CMD Labs - an applied science team within Microsoft Teams - to work on improving transcription accuracy specifically for Japanese, with emphasis on named entities (e.g., names of people/products/teams etc.) by applying existing or novel research and leveraging training, fine tuning, and prompt engineering of speech transformer models, as well as LLMs and audio-enabled foundations models as post-processing and re-scoring modules. Our flagship AI applications for Teams Meetings such Team Copilot, Personal Meeting Copilot and Intelligent Recap are all fully dependent on an accurate meeting transcription as the primary grounding data. Within transcription, the importance of named entities - names of people, projects, products, companies and places - is often the most important, and yet the most challenging for the transcription engine since the names might not be a part of the model's training data. The Research Intern will be onboarded to our evaluation pipeline code which takes as input real meetings that were donated internally, and work to iterate on existing algorithms as well as propose novel solutions to the problem based on recent academic literature. The work done in the Research Internship will contribute towards the algorithm that the engineering team will implement in production. As a secondary priority, given substantial scientific novelty of the approach and results, collaboration on a mutual publication is also possible.

Requirements

• Currently enrolled in a PhD program (or published candidate in MSc program) in Computer Science, Electrical or Computer Engineering, Statistics, or a related field.
• Research Interns are expected to be physically located in their manager's Microsoft worksite location for the duration of their internship.
• Submit a minimum of two reference letters for this position as well as a cover letter and any relevant work or research samples.

Nice To Haves

• Field of research and experience with Japanese.
• Field of research and publications directly related to transcription or the Audio LLMs.
• Japanese fluency.
• Practical experience in training, fine-tuning, and prompt engineering of transformer models or LLMs.
• Practical Python coding experience leveraging PyTorch or TensorFlow or similar framework.

Responsibilities

• Conduct experiments, create and validate metrics, and develop candidate algorithms to improve the accuracy of transcription of named entities in Japanese and reduce chances of error in downstream LLM-based applications.
• Collaborate closely with CMD Labs researchers and engineers to leverage existing assets, datasets, and ensure results can be leveraged back into the product.
• Embody our culture and values.

Benefits

• Industry leading healthcare
• Educational resources
• Discounts on products and services
• Savings and investments
• Maternity and paternity leave
• Generous time away
• Giving programs
• Opportunities to network and connect

What This Job Offers

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