He built a platform that PM Modi used in parliament. Then turned the internal tool his own data team relied on into a $750M enterprise. Varun Banka has been at the intersection of data and consequence since his final year of university.
In 2012, during his final year of an engineering degree in Singapore, Varun Banka and his co-founder Prukalpa Sankar decided to bootstrap a company with $25,000 scraped together from crowdfunding campaigns, business plan competitions, and grants. No investors. No connections. Just a conviction that data, applied right, could change the world.
That company was SocialCops. Within a few years, its data platforms were tracking 42 Indian government schemes, powering what became known as Disha - India's national data platform. PM Modi used it. Every member of parliament had access to it. India's Ujjwala Yojana scheme, which delivered subsidized cooking gas cylinders to hundreds of millions of rural households, ran on SocialCops data infrastructure. The New York Times called them Global Visionaries. Fortune India put Varun on its 40-Under-40 list at 23 - one of the youngest people ever included.
"We went from a data-for-good company to the metadata layer powering enterprise data teams globally."- Varun Banka
But here's the thing nobody tells you about running a data company: your own data team is a disaster. Files everywhere. No one knows who owns what. A simple question - "where does this number come from?" - takes three Slack threads and two meetings to answer. SocialCops was helping governments govern their data while their own team was drowning in the same chaos they were supposed to solve.
So they built an internal tool. A workspace for data teams to collaborate, discover, document, and understand their data. Nothing existed like it. Once deployed, their own team became 6x more agile. And that's when the idea arrived, quietly but clearly: every data team in the world needs this.
In 2018, Atlan was born. Not as a pivot away from SocialCops - both companies run in parallel - but as a separate bet on the enterprise data market. The pitch was direct: data teams were using the same passive, siloed data catalogs that enterprises had been sold for decades. Atlan would build something different. Active metadata - not a static directory, but a living, breathing intelligence layer that continuously captures lineage, tracks usage, surfaces quality issues, and enables the kind of collaboration that turns data from a liability into a weapon.
The market responded. Cisco. Nasdaq. Unilever. Ralph Lauren. FOX. News Corp. HubSpot. Plaid. These aren't pilot customers - they're production deployments at scale, replacing tools that cost more and do less. Atlan deploys in 4-6 weeks where legacy platforms take 3-9 months. It wins 80% of competitive trials.
By May 2024, GIC - Singapore's sovereign wealth fund - and Meritech Capital led a $105M Series C that valued Atlan at $750M. Revenue had grown 7x in two years. Enterprise sales were up 400% in Q1 2024 alone. This is what happens when a data tool is built by people who have actually used data at scale for something that matters.
Most tech founders start with a market gap. Varun Banka started with a moral conviction. In 2012, while still a student, he and Prukalpa Sankar asked a question that most 22-year-olds aren't asking: why are the people making decisions about the developing world's most pressing problems working with such bad data?
SocialCops was their answer. It started as a crowdsourcing platform for social data - a way for citizens, NGOs, and governments to collect field data and transform it into insight. The approach was simple, the ambition was not. Within a few years, SocialCops was running projects across three continents and had become the data backbone of India's largest government initiatives.
The Ujjwala Yojana project is the one that defines the scale. India's government wanted to deliver subsidized cooking gas cylinders to Below Poverty Line households - an initiative that would eventually reach hundreds of millions of people. SocialCops built the data layer that made it trackable, accountable, and improvable. Real-time. At national scale.
"Data teams deserve the same collaborative tools that software engineers have."- Varun Banka
The paradox was always there, quietly. SocialCops was selling data maturity to governments while quietly battling the same chaos internally. Their own data team - tasked with projects that affected hundreds of millions of people - was spending more time figuring out what data existed and who owned it than actually analyzing it. The internal tooling project that fixed this problem took two years to build. It also became Atlan.
SocialCops continues operating today as the social good wing, even as the founders scale Atlan. The two missions - data for society and data for enterprise - coexist, each informing the other.
The consistent thread across everything Varun has built is a refusal to accept that data has to be hard. Not complicated-hard - just unnecessarily hard. Chaotic-hard. Where-did-this-number-come-from hard.
At SocialCops, the chaos was in government. Agencies with conflicting datasets, no lineage, no accountability. At Atlan, the chaos is in enterprise. Data teams that are brilliant but can't answer a simple question without three days of archaeology. In both cases, the solution Varun built was the same: give people context.
Atlan calls it "active metadata." The phrase is deliberate. Passive metadata - what traditional data catalogs offered - was a snapshot. A directory. Useful, but not alive. Active metadata is a continuous capture of everything happening to your data: who touched it, what changed, where it goes, what it means to the business, how trustworthy it is. Varun's insight was that metadata wasn't a documentation problem. It was a context problem. And context is what intelligence runs on.
"Passive, siloed data catalogs are dead. Active metadata is the future of data governance."- Varun Banka
The AI angle sharpens this significantly. As enterprises deploy AI across their data stacks, the question isn't just "what data do we have?" It's "can we trust this data? Does our AI system understand what this column actually means? Who owns accountability for this decision?" Those questions all require context. Atlan's bet - and Varun's bet - is that the "context layer" will become the most important piece of infrastructure in the AI-native enterprise.
It's a big bet. GIC and Meritech Capital funded it to the tune of $105M. The Fortune 500 is showing up. And the kid from Ranchi who bootstrapped his first company with $25,000 is building the control plane for how the world's largest organizations understand their most important asset.