On July 19, 2021, Jiten Vaidya announced a change at PlanetScale. The company he had co-founded would have a new chief executive: Sam Lambert, who had joined as chief product officer nine months earlier. Vaidya would become chief strategy officer and remain on the board. The announcement ran to a few short paragraphs. For a business built around moving complicated things without interrupting service, it was an appropriately economical piece of writing.
The timing gives the decision its interest. PlanetScale had launched a new developer-focused database in May and announced a $30 million funding round in June. Vaidya was handing over the job during a period of product change and fresh investment. He praised Lambert’s hiring, product vision and execution, describing a close working partnership. There was no grand farewell. He expected to keep working.
That handoff is a useful place to begin with Vaidya, because a list of employers can obscure the choices between them. Google, YouTube, Dropbox, the US Digital Service, PlanetScale: the names suggest a career devoted to infrastructure. Look more closely and another concern keeps appearing. Who carries the complexity? Who has to understand every awkward detail? And when can someone else take the wheel?
A taste for evidence, and cat videos
Vaidya attended the Indian Institute of Technology Bombay from 1984 to 1988. His degree was in chemical engineering, an early turn in a career that would become closely associated with software. His professional path included backend and infrastructure work at Google and YouTube, followed by Dropbox and government service. These were different institutions, with different customers, operating under very different kinds of scrutiny.
An old Dropbox biography provides a less ceremonial introduction. It described his previous employer as a search engine company and YouTube as its purveyor of cat videos. He called himself an empiricist, a reader and a walker of trails. This is a pleasingly modest vocabulary for someone whose working day involved keeping other people’s enormous systems upright. Even the internet’s kittens require an operations department.
At YouTube, Vaidya managed the database administration and site reliability engineering teams that operated Vitess. The distinction matters: Sugu Sougoumarane was one of the project’s co-creators; Vaidya brought experience leading the teams responsible for running it. Their later partnership at PlanetScale joined the design of a system to the daily practicalities of keeping it useful.
In 2012, Vaidya left Google for Endorse, a startup subsequently acquired by Dropbox. He later spent four months with the US Digital Service in Washington, working on veterans’ administration problems. The move back toward hands-on engineering appealed to him. His career had room for managing teams, writing software and a short spell inside government, rather than a single uninterrupted climb through corporate titles.
When the application becomes the baggage handler
At YouTube, the difficulty was growth. A database that works for today’s audience can become tomorrow’s constraint. The team added replicas to serve reads, then split data across separate databases as demands increased. That splitting is called sharding. It expands capacity, but it introduces a new question every time the application needs information: which piece holds the answer?
YouTube initially made the application answer that question. Its code contained the logic for directing queries to the right shard. As the arrangement became more complicated, product engineers had more infrastructure detail to carry around. A feature that looked simple from the outside could arrive with a surprising quantity of database luggage.
Vitess, created in 2010, moved that routing work into a layer between the application and MySQL. An application could deal with a logical database while the machinery underneath organized multiple pieces. The approach preserved the familiar database and moved the coordination elsewhere. It also made database management itself something the team could automate.
By 2017, people were asking whether there was a commercial organization they could pay for Vitess support. Vaidya and Sougoumarane founded PlanetScale in 2018. The demand was quite concrete: organizations wanted access to the technology and someone accountable for helping them use it. An open repository can supply code; a customer with a difficult deployment also wants a person who will answer.
The company’s early commercial argument also depended on what customers could keep. MySQL came with established tools, familiar SQL and years of experience inside engineering teams. Scaling it offered a path that preserved that investment. PlanetScale’s pitch included running across regions and cloud providers, giving organizations choices about where their data lived. The attraction was practical: a larger system could still make use of knowledge and software the customer already had, rather than demand an entirely fresh start.
Vitess entered the Cloud Native Computing Foundation’s incubation program in February 2018 and graduated in November 2019. It had an existence beyond the startup. For PlanetScale, that meant building a business around shared infrastructure while offering something customers could distinguish from downloading the software themselves. The company would have to make expertise accessible.
A promise ought to survive a power cut
Vaidya’s technical writing shows what he considered worth explaining. In April 2020, he took on the standard way of teaching database transactions: moving money from account A to account B. The example was familiar, but he wanted readers to think about newer applications too. Correctness has consequences well beyond a bank ledger.
His examples included changing access to a Slack channel and sending a colleague a message. A system says an operation succeeded; the user reasonably believes it. If a failure then erases the change, the confirmation has become misleading. Durability, one of the guarantees grouped under the acronym ACID, concerns whether committed information survives a crash or power failure.
It is a revealing choice of explanation. Vaidya approached the database through the ordinary person waiting for it to keep a promise. The machinery might be distributed across servers, but the experience arrives as something small and definite: a message sent, a permission changed, a confirmation received. Technical correctness becomes a matter of everyday trust.
He explored a related problem in a 2020 conversation with engineer Anthony Yeh: keeping MySQL available across clouds. YouTube’s operators had wanted resilience against losing an entire data center. Other organizations wanted protection against a cloud provider’s failure. Moving from one machine to a fleet of machines widens the list of things that can go wrong, and widens the work required to recover.
Making expertise disappear into the product
PlanetScale’s May 2021 launch changed the emphasis. A hosted version of Vitess had been its foundation. The new product concentrated on developers’ workflows, with database branching and non-blocking schema changes among its capabilities. Branching gave developers a way to work on changes to a database’s structure before bringing those changes into production.
The underlying observation was familiar to the team: a startup chooses a database to get moving, then spends later years dealing with the consequences of that choice. The product aimed to make those early decisions less punishing. Database changes should fit into the way developers already worked, while the infrastructure retained room to grow.
“We want to be your DBA team.”
Jiten Vaidya, June 2021
Vaidya’s description of the ambition was unusually direct. Running Vitess could be difficult; PlanetScale wanted to take that work off developers’ hands. The funding was meant to help bring the new product to market. Expertise would still be needed, but a customer would buy access to it through a service rather than assemble an equivalent team before getting started.
Insight Partners led the Series B, with a16z and SignalFire participating. Total funding reached $55 million. The release described a product intended for a single developer getting started as well as a growing organization. That combination explains the commercial challenge: make the first encounter approachable while retaining the operational knowledge that justified the company’s existence.
Lambert had helped shape that product direction. When Vaidya announced the leadership transition the following month, he credited Lambert with building the team and executing the vision. The title followed work already happening. A founder had made room for a colleague whose responsibilities and influence had expanded with the product.
The next person at the table
Since early 2023, Vaidya has worked part time as an advisor, investor and executive coach. He lists relationships with companies including Rocicorp, Replay, Metaphor, Responsive, Requestly, Spenza and Unlogged, and describes himself as a small investor in and advisor to Ola Krutrim. The mix carries his work between the United States and India.
Namespace also names him among its investors and advisors, alongside people with backgrounds in developer infrastructure and software companies. Its concern is the time developers lose waiting for builds, tests and other infrastructure. That is familiar territory for someone who spent years thinking about the work that gets between an engineer and the feature they intend to ship.
Vaidya describes his coaching role as offering CEOs a sounding board and a listener. His own experience gives that offer some specificity: he has operated systems, managed teams, founded a company, raised money and handed over its leadership. A conversation can draw on all of those roles without requiring him to occupy the executive chair again.
Outside work, he likes hiking, reading and cooking for friends and family. The reader and trail walker from the old Dropbox biography are still recognizable. After years spent helping organizations distribute the load across machines, his work now includes sitting with the person who has to decide what to do next. The useful contribution may be a technical answer. Sometimes it may be the space to finish a thought.