Hi, I'm
Senior Software Engineer at Snowflake
PhD in CS from University of Washington · SQL compiler & query optimizer · Database systems researcher
Background
I'm a Computer Science PhD graduate from the Paul G. Allen School at the University of Washington, where I worked with advisors Bill Howe and Ed Lazowska on generalized SQL workload analytics.
My research lies at the intersection of database systems and machine learning — exploring how NLP techniques can be applied to SQL to enable smarter workload management, query optimization, and data discovery.
Currently, I work as a Software Engineer at Snowflake, where I work on the SQL compiler and query optimizer, helping make cloud-scale analytics faster and more intelligent.
Before my PhD, I was an engineer at Microsoft India Development Center, where I helped build Azure Site Recovery and MOHORO (the precursor to Azure RemoteApp). I completed my undergraduate degree in Computer Science at BITS Pilani.
Snowflake — SQL compiler & query optimizer
University of Washington (Paul G. Allen School)
SQL workload analytics, query representation learning
BITS Pilani
Career
Working on the SQL compiler and query optimizer. Delivered Snowtrail — a production query testing system that replays and compares query behavior across compiler versions. Improving query plan quality and optimizer correctness at cloud scale.
Developed Querc, a system for database-agnostic workload management using learned query representations. Investigated NLP techniques (word2vec, paragraph vectors) applied to SQL. Maintained and analyzed the SQLShare DB-as-a-Service platform, publishing findings at SIGMOD 2016, CIDR 2019, and other top venues.
Introduced support for dynamic types and JavaScript development to Project Orleans, Microsoft's distributed actor framework — enabling a new class of JavaScript-based distributed applications on .NET.
Built HyperV-based Disaster Recovery as a Service (now Azure Site Recovery). Was part of the team that created MOHORO, a scalable pay-by-usage Desktop-as-a-Service on Azure that served as the precursor to Azure RemoteApp.
Researched autonomous pattern formation algorithms for asynchronous swarm robots without agreement on chirality. The resulting work was published in a peer-reviewed paper on theoretical distributed robotics.
Projects
A system enabling generalized workload management tasks using learned query representations — applied to diverse SQL workloads across different database backends. Published at CIDR 2019.
Read paper
Systematically evaluated NLP embedding techniques — word2vec, paragraph vectors, and graph methods — applied to SQL to enable downstream workload management tasks like query recommendation and anomaly detection.
Read paper
A multi-year SQL-as-a-Service platform used by scientists to share and query datasets. Analyzed real-world query logs to uncover patterns in data cleaning, schema usage, and SQL behavior in the wild. SIGMOD 2016 Reproducibility Award.
Read paper
Contributed to RACO — a query compiler and middleware that optimizes queries across heterogeneous backends (relational, NoSQL, parallel systems), choosing the best execution plan across multiple storage systems.
Learn moreAcademic
Toolkit
Get in touch
I'm always happy to chat about database systems, query optimization, machine learning for databases, or interesting engineering problems. Feel free to reach out via email or connect on social media.
You can also check out my research and code on GitHub.
