Yingjun Wu

Ask most engineers about stream processing and you will hear a sigh. It has long been the domain of specialists - people who stand up clusters, tune obscure parameters, and babysit pipelines that break at 3 a.m. Yingjun Wu spent more than a decade studying exactly this problem. Then he decided the difficulty was not a law of nature. It was a design choice worth undoing.

Wu is the founder and CEO of RisingWave Labs, a San Francisco company he started in 2021. Its product, RisingWave, is a distributed SQL database built for stream processing. The pitch is deceptively plain: if you can write a SQL query, you can work with live data. No dedicated infrastructure team. No specialized language. The complexity stays under the hood, where Wu thinks it belongs.

That conviction did not appear overnight. It was earned across research labs, a PhD, and a stretch inside two of the largest data platforms in the industry.

From Guangzhou to the research bench

Wu's path runs through the academic core of database systems. He completed his undergraduate work at South China University of Technology, where he earned an Excellent Graduate Thesis Award. He went on to a PhD at the National University of Singapore under Kian-Lee Tan, writing a dissertation with a title only a systems person could love: Transaction Management In Multi-Core Main-Memory Database Systems. Along the way he picked up the Dean's Graduate Research Award.

During that period he was also a visiting PhD at Carnegie Mellon University, hosted by Andrew Pavlo - a name that carries weight in the database community. The work was serious and it was recognized. Back in 2013, Wu's research on a stream processing engine called SPRINT won the Best Paper Award at DEBS, a conference devoted to event-based systems. Years before he would build a streaming company, he was already publishing on how to make streams fast.

His publication record reads like a tour of the field's flagship venues: SIGMOD papers in 2016 and 2019 on transaction healing and secondary indexing, a VLDB paper in 2017 on multi-version concurrency control, a USENIX ATC paper on scalable transaction processing with hardware transactional memory. He also served the community, sitting on program committees for SIGMOD, VLDB, and ICDE, and helping organize workshops like AIDB, SMDB, and DEBS.

Inside the machine

Research explains the theory. Industry taught him the pain. Before founding RisingWave, Wu worked as a researcher in the Database Group at IBM Almaden Research Center, where he earned a Manager's Choice Award. He then joined Amazon Web Services as a software engineer on the Redshift team - the group behind one of the most widely used cloud data warehouses in the world.

Those two stops matter. Almaden is where deep database ideas get pressure-tested. Redshift is where they meet millions of real queries and unforgiving customer expectations. Sitting at that intersection, Wu saw the same gap over and over: batch analytics had become approachable, but real-time analytics had not. The tools for streaming were powerful and brutal to operate. If you wanted live data, you paid for it with headcount and heartburn.

Building RisingWave

In 2021 Wu started RisingWave Labs, co-founding it with Eric Fu. The design goal was a streaming system that felt like a database, not a distributed-systems project. RisingWave is PostgreSQL-compatible, which means the SQL, drivers, and tools people already use largely just work. It leans on materialized views that update incrementally as new events arrive, so results stay fresh without full recomputation. And it was cloud-native from the first commit, decoupling compute from storage so teams pay for what they actually use.

The company put four principles at the center of everything it ships.

The market noticed. In 2022 RisingWave raised a $36 million Series A to launch its cloud offering, part of more than $40 million in total funding. The open-source project drew thousands of GitHub stars, and Wu assembled a team stocked with engineers who had worked at AWS Redshift, Microsoft Azure, Snowflake, LinkedIn, and Uber. In 2023 the company shipped RisingWave 1.0 and Wu took the stage at QCon San Francisco. By early 2024 he was on Software Engineering Radio explaining, to a wide audience of developers, what a streaming database actually is.

How the idea flows

Strip away the jargon and RisingWave does something intuitive. Events pour in from sources like Kafka and change-data-capture streams. RisingWave keeps SQL queries continuously up to date as those events land. The results feed dashboards, applications, and increasingly a streaming data lakehouse built on open standards like Apache Iceberg.

The use cases are the kind that only work if the data is genuinely live: fraud detection that flags a transaction before it clears, dynamic pricing that shifts with demand, real-time inventory, personalized recommendations, monitoring and alerting. In each, the difference between a few seconds and a few minutes is the difference between acting and reacting.

The researcher-founder

There is a well-worn belief that academics make awkward founders - too fond of correctness, too slow to ship. Wu's arc pushes back on that. His research obsession with consistency became a product feature. His years watching real workloads at Redshift became a customer thesis. The company he built is essentially the system he wished had existed while he was doing the hard version by hand.

He keeps writing, too. Wu publishes on Medium and Cloud Data Insights, and remains an active voice in the streaming and database community. The through-line across the papers, the posts, and the product is the same stubborn idea: powerful data infrastructure should not require a priesthood to operate.

For a category that has spent years feeling out of reach, that is a quietly radical goal - and Wu has staked a company on delivering it.