CV

Principal Engineering Manager

Microsoft

• Currently leading agentic savings in Copilot. This includes leading the development of platforms for product tracking and coupons among other saving opportunities.
• Led the addition of the core shopping scenarios into Copilot as the DRI which involved partnering closely with designers, product managers, UXR, engineering, and applied scientists.
• Led the development effort of adding product tracking inside Copilot, enhancing user shopping experience by providing savings opportunities.
• Doubled the growth of product tracking in six months, driving the lead in innovative features for Microsoft's platform.
• Introduced features to extract user's shopping intent, enhancing the overall Shopping experience on Copilot.

Senior Engineering Manager

Microsoft

• Led the development of Microsoft Autos AI Assistant, the first assistant on an autos marketplace using generative models.
• Spearheaded the creation of "My Garage" for Microsoft Autos Marketplace, consolidating user vehicle data from various sources.
• Developed autos entity frameworks for Bing, MSN, and Microsoft Autos Marketplace.

Data & Applied Scientist II

Microsoft

• Led the fine-tuning of machine learning model for product information extraction on edge.
• Proposed graph representation capturing relations between entities and news articles.
• Reduced resource usage for entity matching workflows by 50% through optimization techniques.
• Implemented Bing Satori entity matching stack on Spark platform.

Data & Applied Scientist Intern

Microsoft

• Designed and implemented a conflation update component to accelerate entity matching changes in the knowledge graph.
• Demonstrated the value of enriching source graphs post RDF triples generation for Satori.
• Developed guidelines for optimizing RDF graph generation for the Entity Platform and Conflation teams at Microsoft.

Research Intern

Microsoft Research

• Conducted experiments to increase Bing dependency parser accuracy against state-of-the-art parsers.
• Validated impact of Entity Linking on Named Entities accuracy.
• Conducted experiments to validate impact of specific query patterns for high-quality training data.

Research Intern

Yahoo!

• Developed a recommendation system for local businesses.

Research Engineer

IBM

• Tag Recommendation Technique: Co-invented and patented a k-partite graph technique for tag recommendation during resource bookmarking.
• Spam Detection: Proposed a set of features that identify spam bookmarks in social media and evaluated their feasibility across various machine learning algorithms.
• Extracting relations from biomedical text: Developed a system that automatically extracts interactions between entities such as proteins, chemicals, and diseases from biomedical text with preliminary focus on protein-protein interactions.
• Machine Assisted Translation: Developed a system that provides suggestions when translating documents. The system uses continuous space language models for statistical machine translation.
• Unstructured Information Management Platform: Participated in joint development with Watson lab in Yorktown to develop a system that analyzes large volumes of unstructured textual information from the news domain to discover, organize, and deliver relevant knowledge to end-users.

PhD in Computer Science

Purdue University