Rachit Kataria and the price of a missing name

Rachit Kataria stopped two weeks of engineering work to save an account. A customer wanted a fix; his team built it. The account remained in danger. Somewhere between the request and the release, the important detail had escaped everyone: the person making the decision had changed.

For an engineer, this is an especially irritating kind of failure. The work can be correct, delivered, and beside the point. Kataria was a product tech lead at AtoB, the fleet-card startup. The experience would help give him a new problem to work on, one that required understanding who had the power to say yes.

Today he is co-founder and CEO of Centralize, which builds relationship maps for enterprise revenue teams. His career runs through Apple internships, Facebook Shops, and Y Combinator. But the thread worth following is smaller than those names: his persistent interest in the people a system forgets, and the friendship that became his company.

The customer had changed. The team’s picture hadn’t.

In Kataria’s later account of the AtoB episode, the customer represented a $300,000 account. Its Head of Fleet had departed months earlier. The sales representative who originally started the deal had also left. A new leader appeared with unfamiliar questions, while useful context sat inside an eight-month-old Gong recording that nobody knew to consult.

He calls this “relationship amnesia.” It is an apt description of a peculiar office condition: an organization can possess the information and still behave as though it has forgotten everything. The recording exists. The inbox exists. The person who knows which one matters has moved on.

That distinction gives his company a concrete purpose. A contact record can tell you a name and title. A useful account map also needs to show a departure, an overlooked decision-maker, or a relationship that has gone quiet. Otherwise a team may spend its time answering a question whose owner no longer holds the budget.

Before the account map, an Apple Watch

Kataria grew up in the Bay Area, with engineer parents in San Jose. He switched into computer science before his first classes at the University of Southern California, initially intending to try it for a semester. The experiment acquired momentum.

At HackSC in 2015, his first hackathon, his team built Quark, an Apple TV application that let people manipulate video and audio. A participant could swipe to distort video; QR codes brought instruments into the mix. The project’s inspiration came from teammate Hannah Criswell, who combined computer science with flute performance and an interest in videography.

The team built in Swift. Three of its four members were new to iOS development. Apple TV’s two-device connection limit proved an obstacle, which the project write-up greeted with a suitably exasperated complaint about Apple TV. The software won awards for visual media, an Apple platform app, and newcomers.

There was also an Apple Watch. In 2018, Kataria was still wearing the prize he had won. More consequentially, the hackathon introduced him to Apple recruiters and helped lead to his 2016 internship in Cupertino. A weekend project had produced both a watch and a professional introduction. One of those is easier to put on a wrist.

By 2018, he had been selected for the Kleiner Perkins fellowship: 52 students chosen from nearly 2,500 applicants. His placement was at Shape Security. He was already thinking about the people in his cohort as future collaborators, well before he had a company to recruit for.

“Having the opportunity to learn from CEOs at leading startups is absolutely priceless.”

Rachit Kataria, 2018

A roommate with a very long runway

At USC, Kataria met William Wang while they were applying to Spark SC, the student entrepreneurship organization. They became friends through building things and drinking boba. By 2019, they were roommates pursuing master’s degrees in computer science and discussing whether they might eventually start a company together.

Kataria later described how little some parts of that arrangement had changed. There were still late conversations on the couch, shared jokes, and homemade videos. The subjects had moved from internship adventures to customer instructions. A private vocabulary was becoming feature names and trademark applications. The couch had acquired a second occupation as Wang’s bed.

Their first relationship project was Flock. After graduating, they felt friends drifting away and built an app to help people keep close connections. It was a personal problem with an obvious emotional stake. Their early customer was anyone who had discovered that affection alone does a poor job of organizing a calendar.

Eventually they stopped working on Flock and turned toward revenue teams. The continuity is revealing. The setting changed from friendship to business, but both projects asked how people could retain context about relationships that mattered. Centralize gave that question a commercial setting, where losing the thread could mean losing an account.

Learning what survives a handoff

Before Centralize, Kataria worked on products with very different scales and customers. His résumé records a senior software engineering role at Facebook from January 2020 through April 2022, followed by a tech lead and staff engineering role at AtoB beginning in May 2022.

At Facebook Shops, he helped build the mobile storefront and tools that enabled businesses to set up shop. He also led an infrastructure effort supporting multiple Shops per business, coordinating across eight teams. These were engineering assignments with organizational consequences: a useful product still had to accommodate how actual businesses operated.

At AtoB, his work included partner integrations, discounts, fleet administration, and fraud prevention. The job put him near the commercial conversations around a technical product. The endangered account exposed a limit to what another release could accomplish when the team’s understanding of the customer was already out of date.

Wang brought a different product background, having started Slack’s prototyping team and led the initial launch of Huddles. The two founders arrived at Centralize with experience turning ideas into working products. Enterprise selling added a harder-to-specify requirement: keeping track of who trusted whom, and how that changed between conversations.

Putting the people back on the map

Centralize was founded in 2023 and joined Y Combinator’s Winter 2024 batch. Its software draws on CRM records, emails, and calls to assemble account maps. Sellers can see people they know, people they have yet to meet, and possible introductions through the wider team.

There is a practical difference between collecting contacts and understanding a buying committee. Ten conversations can still leave the budget holder untouched. A friendly user may care deeply about a feature without having authority to purchase it. A map becomes useful when it helps a seller recognize that difference before the next meeting.

Centralize’s current product also brings the account team onto the same view, with shared tasks and account context. Its AI assistant, Centra, is designed to answer questions about budget ownership and routes to senior executives. The aspiration is a team that can discuss its next move without first reconstructing the last six months.

Kataria has taken that argument into conversations with sales practitioners. In March 2025, he and Wang joined Peter Ahn on Decoding Sales to discuss relationship-led selling, outbound approaches, and hiring their first seller. Another conversation with Andrew McGuire examined the relationships available through executives, investors, and advisers. His engineering background had become a way into sales strategy.

A company built around introductions

On October 9, 2025, Salesforce Ventures welcomed Centralize to its portfolio. Its investment thesis emphasized the difficulty of finding real champions and navigating decision-makers when context is dispersed across messages and notes. The founders’ experience working alongside go-to-market teams was part of the appeal.

In July 2026, Centralize announced a $15 million Series A led by NEA, bringing total funding to $19 million. It also introduced a free tier for individual sellers. The pairing made the distribution plan tangible: let a representative build a map, use it on a real account, and bring colleagues into the work.

In September 2026, the founders appeared with YC’s Diana Hu to explain the product and its underlying relationship graph. Their ambition had grown from keeping friends connected to preserving a company’s memory of its customers. The question remained recognizably human: who knows the person we need to speak to?

The next conversation

Kataria’s public descriptions of hiring offer a view of how he wants Centralize to work. He interviews every candidate. Engineers can use AI coding tools during a five-hour feature-building exercise. He pays attention to clarifying questions, reactions to disagreement, and whether candidates start with the customer’s problem.

Those choices put behavior ahead of a polished account of behavior. A candidate can describe being adaptable; a disagreement gives them an opportunity to demonstrate it. A candidate can list tools; a bounded building task shows how they use them. For someone whose company grew from an expensive gap in shared understanding, the attention to communication is easy to follow.

There is a pleasing consistency to the story. An early hackathon brought Kataria into contact with recruiters. A student organization brought him a co-founder. A fellowship offered people he hoped to call when he started something. The account at AtoB showed what could happen when such connections disappeared from view. He has spent years learning how much can depend on remembering the right person. Now he is building somewhere for a team to put that knowledge.