Aug 2012 – Jun 2021 · 8 yrs 10 mos · Greater Seattle Area

• Organizational Leadership
• Part of ThoughtSpot founding/leadership team, helping company grow (600+ now), in establishing our engineering practices and culture, key role in defining & realizing our company product, vision & road map.
• Working directly with customers (including several fortune 500), sales team during all phases of sales cycle: POC, pre-sales & post sales to help customers succeed.
• Established & heading our Seattle area office, our first development organization outside Bay Area.
• Technical Leadership
• 1. Driving SpotIQ. SpotIQ bridges the gap between search and AI driven analytics to bring the power of both paradigms at the hands of end users. Check https://thoughtspot.com/spotiq for SpotIQ details!
• 2. Driving backend platform. Conceptualized & developed Falcon, our high performance, scale out, distributed database from scratch for purposes of enabling instant search experience over TBs of enterprise data with consistency and transactional guarantee of database.
• a) Database system supports complex ad-hoc sql queries (sub-second over billions of rows), and updates (non blocking, multi-billion rows/hr) at extremely high performance. Auto-scaling when nodes are added and vice-versa, resiliency to node, process restarts , hands off administration - critical operational aspects of the system.
• b) Built the data ingestion pipeline to enable a TB/hr scale-out load speeds within the platform. Key - support multiple parallel super optimized loads to a single table with ACID guarantees.
• c) Leveraging the stack, further built the ELT framework within TS to enable data transformations at our scale and performance. Primarily for customers who had multi-TBs of data, needed to perform data manipulations for analytics and couldn't get it done at traditional ETL systems.
• 3. Working with support/operations team in building self-support capabilities in the system to reduce support cost in the field. This is critical to enable us to scale with customer base.

Apr 2011 – Aug 2012 · 1 yr 4 mos · San Francisco Bay Area

• Ads backend system provided the core and mission critical infrastructure system to enable analysis of Ads related data. The system was distributed and supported massive scalability in terms of "extremely large" data sizes, very high update rates, fault tolerance coupled with strict transactional semantics and with very high querying performance both in terms of latency and throughput.
• 1. Driving system architecture & development to lower organizational cost by improving resource utilization while maintaining the high throughput and low latency required of the system.
• 2. Design & develop the next generation analytics system for 100x scale.

Feb 1999 – Apr 2011 · 12 yrs 2 mos · Greater Seattle Area

• 1. Built & led Conference Management service team. It is a scale out, customizable peer reviewing web based service (http://cmt.research.microsoft.com/cmt).
• As manager of the team, built the team from ground up, people management (zero attrition rate)
• As chief architect. led all aspects of the service end to end - requirement gathering, software architecture, development, support, operations and release management.
• Strongly focused on automation to keep organizational costs low (one support person across 1000+ programs, growing 250K+ active userbase).
• Drove external adoption from fewer than 100 to 1000+ programs within 3 years.
• Evangelized service within Microsoft to help run several key programs on the service platform. Achieved by making service very versatile & customizable.
• 2. Led the design and development of Database Tuning Advisor in SQL Server to automate database performance tuning.
• Significantly improved scope and scalability (10x-100x speed ups) compared to its predecessor.
• Scaled to much more complex environments - 10-100x times number of tables, queries, query complexity.
• Driving the work to automate performance tuning in SQL/Cloud scenarios.
• In recognition of significant long term impact of this work in both academic and industry, AutoAdmin was awarded the ten year award in VLDB 2007.
• Developed key technologies in bridging the gap between structured and unstructured data at scale with significant impact in both research & industry specifically for ads, search.
• 3. Developed Query Portals a system to demonstrate how entities can be used to provide a much richer search experience & its relevance to ads at scale.
• Automated entity extraction from web pages (billions of pages, few ms overhead per page).
• 4. Developed DBXplorer to search database using simplicity of keywords.
• In recognition of significant long term impact of this work in both academic and industry, it was awarded most influential work in ICDE 2012.