An app can look finished long before it is useful. The buttons respond. The screens glide. Then somebody tries to sign in, save something, or use it on another phone, and the work behind the work makes its entrance. This was the territory Ilya Sukhar chose for Parse: all the machinery a developer needs before the interesting part can begin.
There is a certain comic unfairness to infrastructure. Get it wrong and everyone notices. Get it right and people congratulate the app. Sukhar helped build a business around accepting that bargain. Parse offered shared backend services so that teams making mobile applications could stop constructing the same supporting systems for themselves.
His career has since taken him through Facebook, Y Combinator and venture capital. Today he invests at Matrix, with a focus on AI and infrastructure. Yet the thread connecting those jobs is easier to see in the smaller things: a browser extension that lets you keep your hands on the keyboard, a website that gathers housing data, an investing essay that turns a recurring decision into a rule. Again and again, he has found something unnecessarily cumbersome and tried to remove a few steps.
The engineers get a vote
Before the acquisition or the conference stages, there was an immigrant family in San Francisco. Sukhar came to the United States from the Soviet Union as a child and grew up in the Bay Area. His parents initially cleaned and painted homes. They later became entrepreneurs, both programmers themselves. In his account of that upbringing, building things was close at hand: he made games and shared them over IRC.
That detail gives his engineering career a more useful starting point than a funding round. A game made for friends has an audience who can tell you immediately whether it works. Distribution is personal. Feedback is difficult to escape. The pleasure comes from getting something into another person’s hands.
At Cornell, he studied computer science and operations research, earning his undergraduate degree in 2007 and a master’s in 2008. There was also an interest in the consequences of technical choices. In 2005, as president of the Undergraduate Society for Intellectual Property, he helped bring a discussion of copyright and digital restrictions to campus.
His observation then was unusually direct: “engineers are really going to be the ones making the decisions, almost more so than the managers.” The issue was how systems govern what users can do. Years before Parse, he was already thinking about the power that comes with deciding how software works.
“Engineers are really going to be the ones making the decisions, almost more so than the managers.”
Ilya Sukhar, at Cornell, 2005
Cornell’s surviving course pages supply a less ceremonial glimpse of him. In autumn 2007, he was a teaching assistant for Foundations of Artificial Intelligence. His Friday office hour followed the final lecture of the week. Today’s AI investor once had a very literal appointment with students trying to understand the subject.
A business beneath the screen
Sukhar began his startup career as an early engineer at Ooyala, the video platform company. He then became the first employee at Etacts, a contact-management startup that Salesforce acquired. These were opportunities to work close to products and the people building them, before he had a company of his own.
He has described those early startup years as involving layoffs, pivots and difficult fundraising during the 2008 financial crash. The experience matters because the later biography can look deceptively tidy: engineering jobs, startup, acquisition, investing. His own account contains rather more turbulence between the nouns.
Parse emerged in 2011 with Sukhar, Kevin Lacker, James Yu and Tikhon Bernstam as co-founders. The company joined Y Combinator’s summer batch. Its proposition was practical. Mobile teams could use common services for storing information, managing users and sending notifications, instead of assembling and maintaining every component themselves.
Shared plumbing. More room to build the product.
The attraction of that arrangement is easy to underestimate if you have never been the person maintaining the plumbing. A customer rarely chooses an app because its developer wrote an original login system. Yet the system still has to work. Parse let small teams buy access to capabilities that would otherwise consume their time.
This was also a particular moment in software. Phones had become a place where businesses needed to be, and the technical work extended beyond whatever appeared on the screen. A service dedicated to those common requirements had an audience well beyond one type of application. Sukhar’s company occupied the useful space between a developer’s idea and the infrastructure required to deliver it.
Facebook comes downstairs
In April 2013, Facebook acquired Parse. The reported price was $85 million; the companies did not publicly disclose the terms. For Sukhar, it meant continuing to build the product inside an organization whose reach was considerably larger than his own startup’s.
The acquisition brought two different ambitions together. Parse wanted to make application development easier. Facebook wanted a deeper relationship with the developers building the mobile world around it. Those aims could reinforce each other, although the needs of an infrastructure customer and a large consumer platform would never be identical.
Sukhar stayed to run Parse and took on other platform projects. He eventually led Facebook’s developer products, including early work on a developer platform for Messenger. His public appearances from those years chart the change in scale: keynotes with Mark Zuckerberg, a presentation with Amazon’s Werner Vogels, and conversations about what developers could build next.
He left Facebook in 2015. The following year, Facebook announced that its hosted Parse service would close in January 2017. An open source version allowed developers to continue on infrastructure they controlled. The chronology is worth keeping straight: Sukhar had already departed when the shutdown was announced.
The service’s closure complicates any comfortable acquisition ending. For a founder, a sale is a major event. For a developer depending on a service, the important question remains what happens to the application tomorrow. Parse belongs to both histories, and they do not end on the same date.
The open source project continues today. Parse Server can run independently or alongside an existing web application. Its continued development gives the original idea a life beyond the hosted business: other people can still use and change the software, even after the corporate arrangement that once supported it has gone.
Back to the first few people
In December 2014, Y Combinator announced Sukhar as a part-time partner. He was returning to a program he had experienced from several directions: Etacts employee, Parse founder, and then adviser. The sequence supplied a useful range of vantage points on the earliest stage of a company.
He joined Matrix as a general partner in 2016. Its current team page associates him with companies including Fivetran, Chameleon, Flock Safety, Mashgin, Parabola and Sieve. His personal angel investments include Scale AI, Applied Intuition, GitLab and Airtable. These are separate parts of his investing history, rather than a single portfolio assembled under one firm.
In a March 2025 podcast conversation with Boris Epstein and Alexis Munger, he revisited the transition from engineering to founding and investing. The discussion included technical founders’ ability to iterate, persistence, and how early teams handle disagreement. Those subjects are close to the daily work of making a company function before its story has become presentable.
There is continuity here with the earlier jobs. At Etacts he had been employee number one. At Parse he had helped make the product and run the company. Investing lets him return to teams facing decisions whose consequences are still uncertain. The titles have changed; the first few people in a room remain part of the attraction.
The browser and the neighbourhood
Sukhar’s side projects offer a smaller, more intimate account of his interests. Vimium, which he lists on his personal site, brings Vim-style keyboard control to the browser. Its credits name him and Phil Crosby as authors and maintainers. Users can move through pages and open links with keyboard commands. Even a reach for the mouse can become a design problem.
Then there is piedmont.house, his site collecting residential real-estate data for Piedmont, California. It presents sales records in a searchable table. This is a different scale of usefulness from a platform serving mobile developers, but the resemblance is there: gather scattered material and make it easier for somebody to work with.
Neither project needs an elaborate explanation to communicate what it does. One changes how you navigate. The other helps you examine a local market. Both are reminders that a person who has spent years around acquisitions and venture financing can still find a satisfying problem in a keyboard shortcut or a table of house sales.
His essay on angel-investment check sizes applies a similar instinct to his own decisions. The advice is to choose a consistent amount rather than repeatedly improvising. He examines the effects on negotiation, judgment, regret and diversification. It is an engineer’s treatment of an untidy human activity: reduce a variable, then pay closer attention to the decision that remains.
The through line in Sukhar’s work is a preference for making the next action easier to take. At Parse, that action was building an app. With Vimium, it might be opening a link. At Matrix, it is helping another team get started. The scale varies considerably. There is still plenty to do after the unnecessary work has been removed.