Jan 2024 – Present · 2 yrs 4 mos

• As a Group Engineering Manager, I lead multiple teams focused on driving innovation and delivering cutting-edge solutions across Microsoft's Copilot initiatives. Our work spans across Copilot for Search, the Copilot Ecosystem, advanced Debugging platforms, and experimentation capabilities.
• Copilot for Search
• Our team has been pivotal in ensuring that Copilots provide high-quality, low-latency responses by intelligently selecting the best portions of enterprise documents. We helped build semantic search from the ground up for Copilots. In 2025, we're pushing the boundaries further by reducing latencies of our GPU-based models, aiming for end-to-end latencies of under 1 second.
• Copilot Ecosystem
• With 30+ Copilots integrated into key Microsoft products such as Word, Copilot, and Azure AI Studio, my team is driving platform maturity. We've established seamless onboarding processes and best practices that optimize partner experiences. Looking ahead to 2025, we're enhancing transparency around dependencies and building robust communication systems for platform changes. Our observability platform will offer partners end-to-end visibility into availability, latency, and quality metrics, ensuring continuous improvements.
• Debugging Platform
• We successfully launched a V2 experience that offers a streamlined view of telemetry and a side-by-side comparison tool for different experiences. In 2025, we aim to turn this into a flexible platform that empowers engineering teams to build on top of our debugging tools, tailoring them to their specific needs while enhancing collaboration across the organization.
• Experimentation Platform
• Our experimentation platform has evolved from a manual, multi-week process to a more automated and efficient system, reducing experiment cycles to two weeks. In 2025, we are focused on pushing the limits of experimentation agility, with the goal of completing experiments in just 1.5 days and delivering highly accurate scorecards.

Jan 2022 – Present · 4 yrs 4 mos

• Copilot Eval Platform: Successfully led a cross-functional team to deliver a comprehensive evaluation platform in under 3 months, enabling the execution of Sydney experiments with deep insights for Copilot engineers. Achieved high availability and low latency for handling a substantial number of experiments daily.
• Developer UI for M365 Chat Copilot: Orchestrated the enhancement of the main debugging tool for M365 Chat, catering to local and browser debugging needs while introducing crucial features. Drove its adoption across teams, showcasing its value with 7000 daily requests and a strong demand for additional features.
• Responsible AI for M365 Chat Copilot: Spearheaded the development of an AI strategy and integration plan for M365 Chat, collaborating with multiple teams to ensure alignment. Achieved DSB signoff for compliance, successfully transferring ownership and knowledge to another team.
• Search Enhancements: Managed a challenging landscape by prioritizing projects with reduced funding. Oversaw significant improvements in migration projects, achieving parity in .NET migration with latency gains and successful partnerships for latency improvements.
• Search Ranking Platform: Engaged in multiple improvements, including integration enhancements, online performance validation, and serving as a liaison for query and ingestion-related issues, ensuring progress and issue resolution across the stack.
• Various Projects and Initiatives:
• M365 Business Chat: Contributed to grounding data design and delivery, evaluation pipeline setup, RAI definition, and escalation handling.
• Greenland: Managed rollout and issue investigations, achieving significant latency improvements and handling customer escalations effectively.
• AirGap & Semantic Search: Provided liaison support and enabled critical features for these initiatives.
• LSS Platform: Improved on-call experience, resolved pain points, and progressed .NET 6 migration despite challenges.

Aug 2019 – Jan 2022 · 2 yrs 5 mos

• Project Leadership & POC Success:
• Led the completion of the POC for Bing over Azure, showcasing exceptional project management skills and pioneering the successful migration of the index, achieving latency goals and setting a benchmark for other teams.
• Technical Enhancements and Stability Improvements:
• Implemented significant stability improvements in production environments, drastically reducing escalations and DRI workload, leading to more reliable systems.
• Drove advancements in various areas such as Express tier, ingestion improvements, and the WebExpress project, contributing to increased efficiency, stability, and scalability.
• Strategic Planning & Long-Term Initiatives:
• Played a pivotal role in devising long-range plans for Bing's 2 by 4 goal, expanding fersh index without additional resources and strategizing for future platform enhancements.
• Collaborated on diverse projects like FPGA capability integration, fresh index 8X growth, showcasing forward-thinking strategies to meet ambitious goals.
• Technical Innovations & Collaborations:
• Innovated with JumpHash enhancements, presenting novel solutions and initiating productive discussions for efficient caching techniques.
• Collaborated with various teams (News, Recommendation platform) to boost index sizes, enhance platform capabilities, and contribute to Bing's recommendation systems.
• Performance Improvements & Technical Leadership:
• Achieved a 4X latency improvement in calendar search, addressing customer pain points by optimizing code and enhancing search performance across different rings.
• Steered critical projects like Parallel Index and Bing-Tail L2 filter ranker, overseeing architecture, compliance signoffs, and project management, facilitating smoother workflows and enhancing relevance models.
• Architectural Innovations & Capacity Optimization:
• Led the delivery of Fresh index compaction, a crucial architectural change expected to recover significant capacity and improve system efficiency.

Aug 2015 – Aug 2019 · 4 yrs

• My team is responsible for indexing and ranking the 'fresh' corpus of Bing. The biggest challenge is to return best matches over billions of documents within few milliseconds while the fresh documents and their updates are streamed to the index. And of course this has to be done while maximizing all the machines resources. All of this combined with super passionate engineers makes this a great place to work.

Jan 2012 – Aug 2015 · 3 yrs 7 mos

• Redesigned the fresh index platform to allow for more efficient use of resources and helped push the system to 10X. See customized index patent: https://patents.justia.com/patent/20140181122

Dec 2007 – Jan 2012 · 4 yrs 1 mo

• Worked on redesigning the user interface for adCenter and the back end using Model-View-Control to support automated testing. Some of the other accomplishments include creating a consolidated logging mechanism for adCenter and measuring the credit fraud on the system and then finally worked on implementing list of proposals to reduce charge backs. Lastly designed and implemented API's for third party engineers using WCF.