Mar 2024 – Mar 2026 · 2 yrs

• Architected and led Nvidia GPU integration for Apache Spark and RAPIDS on Azure Synapse, enabling large-scale, accelerated compute for ML and Spark workloads. Average observed increase in job performance was around 2x, with optimal workloads seeing 10x improvement.
• Tech lead responsible for re-designing and implementing improvements in my org’s CICD pipelines to bring test coverage from 70% to 95% and reduce pipeline queuing time by 300x by parallelizing across the stack.

Aug 2020 – Mar 2024 · 3 yrs 7 mos

• Working on Big Data technologies and Apache Spark offerings in Azure Synapse Analytics, Azure HDInsight and Microsoft Fabric.
• Designed and implemented a low-latency cluster pooling service powering Azure Fabric 5x5 experience, enabling job start times in seconds and reducing idle resource overhead. Reduced job start-up wait time by 90%
• Implemented and optimized .NET for Apache Spark API layer, improving execution performance and developer productivity for C# data engineers. Open-source project bringing Spark to C# developers.

Jun 2018 – Aug 2020 · 2 yrs 2 mos

• Designed and led the error classification project across entire Azure Synapse and Fabric stack to improve error messaging and debugging experience of customers, resulting in reduced tickets by 47%.
• Collaborated with performance engineering teams to profile workloads, reduce cluster spin-up time, and tune job scheduling for distributed data pipelines, enhance system scalability and resource utilization.

Aug 2017 – Feb 2018 · 6 mos · Phoenix, Arizona Area

• Doing research in Artificial Intelligence - task and motion planning in robotics. Developing Motion Planning algorithms for Fetch robots.

May 2017 – Aug 2017 · 3 mos · San Francisco Bay Area

• Developed an integration testing environment using Docker Compose to spin up and test multiple micro-services with inter-service dependencies, using a single command. Also built a user-friendly application for entering new fee schedules to the database.

Jul 2014 – Apr 2016 · 1 yr 9 mos · Bangalore

• Involved in the modem protocol development for LTE and RCS (Rich Communication Suite) features in the modem like chat, content-sharing, file transfer, social presence etc.
• Implemented functionality of establishing connections, sending and handling HTTP messages.
• Studied different compression algorithms suitable to compress SIP messages, and responsible for the design and implementation of the entire module for compression and decompression of the messages. Implemented the LZJH algorithm for compression and the UDVM (Universal Decompression Virtual Machine) for decompression of SIP messages.

Jan 2014 – Jun 2014 · 5 mos · Bangalore

• Part of the Modem protocol development team in Samsung Research & Development Center, Bangalore. Responsible for developing a SIP (Session Initiation Protocol) parser. Implemented a parser for SIP messages, consisting of decoder and encoder. Improved the look-up time for header fields in SIP message by implementing a hash table that improved complexity to linear time.

May 2012 – Jul 2012 · 2 mos · Barcelona Area, Spain

• Helped develop product called KeyLess Check-In, an innovative solution to property visit automation.
• KeyLess Checkin is a property automation system that allows remote access of the user's property, in a supervised and secure environment, to the user, or prospective tenants. Responsible for developing the web application that deals with user interface and property management.