New 
Senior Applied Scientist
 Microsoft | |
 United States, Washington, Redmond | |
 Jan 31, 2025 | |
|
OverviewAre you passionate about large-scale machine learning? Are you interested in large language models? We are Microsoft AI, Web Data Platform Document Understanding team and we are looking for a Senior Applied Scientist who will work on developing cutting-edge solutions for various document understanding tasks.We build web-scale Index along with deep understanding of the document content to retrieve rich facets, multi-modal semantic view, dense representations while weeding out poor quality pages to power relevant and delightful experiences across products like Copilot, Bing Search Engine, Personalized Recommendations, Ads, windows experience, etc. Document Understanding Team plays a pivotal and central role in understanding various public web information needs of users and building next-generation models to stay ahead of the curve and push the platform capabilities to go hand in hand with modeling improvements. Work in our team is unique as you'll be able to work with industry-leading scales of data (raw and training data), computing resources and best-quality large language models like GPT4, multimodal large language model, etc.In this role, you will be leading projects from idea creation through implementation, experimentation and delivering improvements to real world scenarios, working closely with various partners. We are looking for a passionate and motivated team member who has solid, hands-on experience in transforming business problems into ML problems, collecting high-quality labels, and developing state-of-the-art models to address product challenges and drive value for end users. You will also leverage your software engineering skills and ML expertise in fields such as natural language processing and information retrieval to help create the next generation of text representation and understanding techniques at Microsoft. 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.
ResponsibilitiesLead innovation and development of deep learning models for document understanding and their usage in downstream tasks, e.g., Copilot, Generative Search, Search, Spam, QA and recommendation, etc.Identify opportunities in Web Data space to solve using ML at scale of 100s of Billions of documents.Push the state-of-the-art in those areas through multiple aspects, for example:Defining the problem spaceGathering training data at scaleExploring model design and architectureExploring learning objectives and tasksBuild Feature Generation Algorithms used at Index Generation time.Build Automated Document Understanding Training Pipelines.Guide team members to develop new technologies that lead to solutions that impact real production scenarios in Microsoft.Work closely with various teams in WebXT to understand common needs and build technical roadmap for addressing them. | |