Shaswat Gupta

Shaswat Gupta is a Software Developer at IBM Research - Zurich, working at the intersection of data systems and machine learning. His research centers on query routing for heterogeneous lakehouses - how to route queries across engines like Spark, PrestoDB, DuckDB, and across increasingly heterogeneous hardware, including GPUs, by jointly reasoning about latency, memory, and cloud cost.

Gupta holds an MSc in Computer Science from ETH Zürich, where he specialized in Machine Intelligence and Database Management Systems. His master's thesis (Prof. Gustavo Alonso, Systems@ETHZ) on this query-routing problem received the top grade of 6/6; a related paper was accepted to a SIGMOD 2025 workshop, with an extended version currently under review at VLDB.

Before ETH, Gupta was a Data Scientist at AB-InBev, building machine learning systems for supply chain and commercial analytics. This shaped a conviction that carries into his research: clean abstractions matter, but so do observability, failure modes, and the details that only surface once a distributed system goes into production.

He completed his undergraduate studies at IIT Bombay, graduated top of his batch (Department Rank 1) with several academic awards, while pursuing a parallel interest in journalism and technical writing.