Sep 2021 – Oct 2024 · 3 yrs 1 mo

• Bing is a comprehensive search engine developed by Microsoft, offering a wide range of services beyond traditional web search—including image, video, and map search capabilities. It is available in 105 languages and operates in 238 countries, serving over 100 million daily users worldwide.
• Bing also powers Microsoft Copilot, an AI-driven chat assistant integrated directly into the search experience. Built on the state-of-the-art multimodal large language models, Copilot delivers up-to-date information, creative content, and intelligent assistance across a broad range of tasks.
• Led a team of 15 engineers focused on delivering real-time sports scores for Bing and Copilot.
• Scaled the experience globally to over 100 regions, enabling sports fans worldwide to follow their favorite teams from anywhere.
• Ensured factual accuracy in Copilot responses by minimizing hallucinations and maintaining high data precision.
• Achieved a 5x increase in league and team coverage, with a focus on expanding support across Southeast Asia and Latin America.
• Reduced real-time update latency from 4 minutes to just 30 seconds, significantly enhancing user experience during live games.

Jul 2017 – Sep 2021 · 4 yrs 2 mos

• Worked on the UX Platform for Bing, driving performance and experimentation at scale.
• Improved mobile page load time by 5% globally through targeted optimizations across critical rendering paths, lazy loading, and resource bundling.
• Led org-wide performance compliance initiatives across a 500+ engineer org, developing internal tools, processes and performance regression tracking to ensure all feature rollouts met defined SLAs.
• Acted as the subject matter expert for A/B testing infrastructure, supporting teams in experiment design, metric instrumentation, scorecard analysis, and statistically sound ship criteria using internal experimentation platforms.

Feb 2015 – Jun 2017 · 2 yrs 4 mos

• Worked on customer feedback systems for local business search in Bing, driving quality and responsiveness at scale.
• Increased user feedback volume 10x by redesigning feedback collection flows and optimizing placement and visibility within the local business search experience.
• Reduced feedback resolution time from multiple weeks to under 24 hours by building automated triaging, prioritization, and routing systems for faster issue handling.
• Led cross-functional improvements across UX design, feedback processing pipelines, and user outreach workflows (e.g., automated email follow-ups) to enhance user trust and engagement.

Nov 2013 – Jan 2015 · 1 yr 2 mos · Bengaluru Area, India

• The AMD Power Profiler, part of the AMD uProf suite, is a performance analysis utility designed for developers and system administrators to monitor power consumption and other metrics across AMD hardware.
• Led the development, packaging, and release of Linux Kernel Modules for the AMD Power Profiler toolset.
• Enabled power telemetry and per-process monitoring on AMD platforms, allowing developers to analyze and optimize application-level power usage.

Nov 2008 – Nov 2013 · 5 yrs · Bengaluru Area, India

• ECI Telecom provides network services to various sectors including service providers, energy, transportation, government, and defense.
• They offer an Optimized Multi-Layer Transport platform, carrier-class transport platforms, carrier Ethernet network, and bandwidth management. Their clients in India include major telecom operators like Airtel, Indicom and Reliance Communication.
• Lead engineer working on a new product requirement to enhance the infrastructure for an existing CES based switch series to support new line cards.
• Added support for LDP/MPLS protocols and APS functionality with a 50ms High Availability promise.
• Worked on extensions to existing LDP/MPLS protocols and development of the CES Pseudo Wire features and APS functionality in Control Plane

Mar 2006 – Oct 2008 · 2 yrs 7 mos · Hyderabad Area, India

• reQall was a voice-enabled personal productivity platform that integrated phone, email, text messaging, and instant messaging to act as a mobile memory aid. It captured daily user input and used contextual signals—such as time, location, and calendar events—to proactively deliver relevant reminders and notes. This intelligent “Memory Jogger” experience was designed to be context-aware and dynamically updated, helping users stay organized and productive.
• Founding engineering team member, responsible for designing and building the backend infrastructure powering Web, iPhone, and BlackBerry clients.
• Scaled the backend system to support over 1 million users and 1 million minutes of audio processing, enabling reliable real-time voice-to-text functionality and seamless synchronization across platforms.
• Delivered significant improvements in system performance, reliability, and user satisfaction through architectural enhancements and efficient data processing pipelines.

Jul 2005 – Feb 2006 · 7 mos · Hyderabad Area, India