Suresh Srinivas

Suresh Srinivas has spent most of his career on a single, stubborn problem: how to make the enormous amounts of data that companies collect actually useful. It is a problem that has changed shape many times over two decades, and he has followed it through nearly every major shift in how the world stores and understands information.

He did not start in computer science. At the National Institute of Technology Karnataka - one of India's top engineering colleges, once known as KREC - he chose electronics and communication engineering on purpose. "I wanted to better understand how things work," he has said. That instinct, to look under the hood rather than take a system on faith, would define the rest of his working life.

His first long stretch in industry was at Sylantro Systems, a telecommunications startup where he spent nearly nine years building IP telephony services in C++ and Java, working with in-memory databases and the kind of distributed systems that would later become central to his career. Market crashes and the rise of mobile eventually limited the company's prospects, but the foundational knowledge stuck.

The Hadoop Years

The turning point came at Yahoo. Inspired by Google's published papers on MapReduce and the Google File System, engineers there were trying to solve a web-scale problem that traditional storage could not handle. Srinivas became an early contributor to Apache Hadoop, focusing on HDFS, the distributed file system at its core. The idea was radical for its time: instead of moving data to computation, move computation to where the data already sits, spread across thousands of commodity machines that scale horizontally.

He has been an Apache Hadoop developer, committer, and PMC member at the Apache Software Foundation since 2008 - a credential that puts him among the small group of engineers who shaped the technology that powered the big-data era.

Building a Company Around Open Source

In 2011 he co-founded Hortonworks as a founding engineer and VP of Engineering, with a mission to commercialize Hadoop for enterprise customers. There he learned a lesson that shaped how he thinks about products. Academic-quality open-source code, it turned out, needed enormous additional engineering to be usable in production. His teams built tooling like Ambari for installation and management, integrated Hive for SQL support, and folded in ecosystem projects like Kafka and Storm.

The takeaway was less about features than about restraint. Companies, he came to believe, should prioritize ease of use and developer productivity over piling on capabilities. "Things with the least number of moving parts are easier to operationalize and more reliable," he has said - a principle that reads almost like a personal engineering creed.

Order at Uber Scale

Srinivas later became Chief Architect for Uber's data platform, guiding its systems through a period of massive global growth. There he confronted a different kind of chaos: thousands of microservices, each with its own inconsistent notion of basic concepts like location. His answer was structure. He helped establish roughly 90 standard types for core concepts and pushed schema compliance and end-to-end governance across the organization. Data culture, in his view, has to be approached with discipline - testing, ownership, and quality monitoring - rather than left to ad hoc habits.

OpenMetadata and Collate

That thinking led directly to his current work. In 2021 he founded OpenMetadata, an open-source platform that connects to more than 100 systems across a company's data stack - from databases to dashboards - and stitches them into a single, unified view of metadata. The premise is simple to state and hard to execute: metadata is the key to unlocking data's value, and producers and consumers of data need a shared, vendor-neutral place to collaborate.

OpenMetadata grew into one of the fastest-growing open-source projects in its category, passing 9,800 community members and earning a Bloomberg FOSS Contributor Fund grant along the way. Around it, Srinivas co-founded Collate with fellow open-source data veteran Sriharsha Chintalapani. Collate turns the project into an agentic data intelligence platform, pairing AI-powered workflows with human collaboration to tackle data silos, quality issues, and fragmented tooling.

In July 2025, Collate raised a $10M Series A led by Venrock, with Unusual Ventures and Karman Ventures joining. Ganesh Srinivasan of Venrock, a former Confluent chief product officer, took a board seat and pointed to the founders' track record. By then the company was reporting more than 350% year-over-year growth in annual recurring revenue, with Fortune 500 customers among its users.

Through all of it, the mission on his GitHub profile stays refreshingly plain: "Passionate about improving data usability & experience." Even his approach to hiring reflects the same preference for the concrete over the abstract. Rather than algorithmic puzzles, he likes to hand engineering candidates a real feature to build during interviews, on the theory that actual work reveals more than any whiteboard riddle. It is the same idea that pulled him toward electronics all those years ago - the belief that the best way to understand something is to build it and see how it holds together.