The Long Game
Most people who co-found a data startup come from one of two places: a PhD lab or a Series C engineering org. Vikas Marwaha came from neither. He spent two decades walking into Fortune 500 boardrooms on behalf of Wipro, Headstrong, and Hewlett Packard, listening to the same complaint: we cannot get clean data when we need it. By the time he co-founded Prophecy in 2019, he had logged seven years as General Manager and Global Partner at Wipro serving financial services giants worldwide. He did not have to guess at the problem. He had seen it up close at every major bank, insurer, and healthcare system in the world.
That context matters. Prophecy is not a tool built by engineers for engineers. It is a platform designed by someone who spent decades watching engineers become the bottleneck - the scarce, expensive resource that every data-hungry organization fights over. The insight is simple but not obvious from inside a data team: the bottleneck is not the talent, it is the interface. Make the interface good enough and an analyst can build what used to require a Spark expert.
How Prophecy Works
The platform operates as a data copilot. Users drag and drop in a visual interface - no code required to start. Behind the scenes, Prophecy generates clean, standardized, open-source code: PySpark, Scala, or SQL. The code is not locked in proprietary format. It runs on Databricks, Snowflake, Apache Spark, and Google BigQuery. Data teams get the speed of no-code with the auditability of open code. IT gets governance. Business analysts get independence.
When generative AI arrived, Prophecy evolved again. The January 2025 $47M Series B extension came partly because the company had already shipped an AI layer that goes further: users describe a pipeline in plain language and the AI builds it. Errors get flagged and suggested fixes surfaced automatically. Documentation writes itself. It is the kind of product that makes a 12-month data engineering backlog feel embarrassingly avoidable.
"Despite decades of investment in desktop and platform tooling and aggressive hiring of hard-to-find data engineers, large organizations still come to us with 12-month backlogs."
- Raj Bains, CEO, ProphecyThe Road to Prophecy
Vikas Marwaha grew up in India and completed a Bachelor of Engineering in Computer Science at Guru Nanak Dev University between 1996 and 2000. He began his technology career as a Sales Manager at Satyam Computer Services, one of India's largest IT services firms at the time. He worked there through 2005, building his foundation in enterprise technology sales during the early outsourcing boom.
An MBA from the Indian Institute of Management, Indore - one of India's elite management schools - followed. He graduated in 2005, specializing in Business Administration and Finance. The combination of an engineering undergraduate and an IIM MBA is a specific type of credentialing in Indian tech: it marks someone who can read code and read a P&L.
HP came next as Client Partner for Enterprise Business, then a move to Headstrong as Sales Director - a firm that was later acquired by Tech Mahindra. In 2012, he joined Wipro, where he would spend seven years. As General Manager and Global Partner for Financial Services, he built and maintained relationships with some of the world's largest institutions. It was work that required understanding regulatory complexity, risk appetite, and the specific IT constraints that come with running 100-year-old financial infrastructure.
When he walked away from that role in 2019 to help build a startup, it was not an impulsive bet. It was a calculated move by someone who had spent years identifying exactly where enterprise data was broken - and had a plan to fix it.
Prophecy Funding Journey
Investors include Smith Point Capital, Insight Partners, SignalFire, JPMorgan Chase, HSBC, Databricks Ventures, SignalFire
The Enterprise Problem Nobody Wanted to Solve
Data pipelines are not glamorous. They do not generate TechCrunch think-pieces about disruption. They generate cleaning bills when they break. The data that feeds AI models, dashboards, and quarterly reports has to get from somewhere raw to somewhere trusted - and that journey is full of manual scripts, fragile hand-coded transformations, and undocumented institutional knowledge.
Prophecy's bet is that the visual-plus-code combination unlocks the next class of data worker. An analyst who understands the business logic does not need to learn Spark syntax to build a reliable pipeline. An engineer who was spending 60% of their time on boilerplate now does not have to. The generated code is open, auditable, and runs anywhere - which removes the biggest objection enterprise IT teams have to no-code tools: vendor lock-in.
The results are starting to appear in notable places. Aetion, a healthcare analytics company, used Prophecy to validate over 500 million patient records. HSBC and JPMorgan Chase are not just investors - they are customers. The Texas Rangers baseball team uses it for analytics. These are not adjacent markets. They are a deliberate signal that the platform generalizes across every industry that runs on data.
The Co-Founder Triangle
Prophecy was built by three co-founders with complementary angles. Raj Bains, who serves as CEO, came from Microsoft and Nvidia with deep experience in hardware optimization and AI infrastructure. He had managed Apache Hive and understood the data engineering stack from the inside. Maciej Szpakowski brought engineering depth on the platform side. And Vikas Marwaha brought twenty years of knowing exactly who would buy it and why.
That split - technical founder plus enterprise operator - is a classic configuration for B2B infrastructure companies. What made it work at Prophecy was that Marwaha's enterprise experience was not generic. It was specific to the customers they were targeting: global financial institutions, large healthcare systems, and Fortune 500 technology companies. He did not need to learn the market after founding. He had lived it.
Prophecy v4 and the AI Agent Era
In February 2026, Prophecy launched v4. The release marks a meaningful architectural shift: AI agents that take business intent - described in natural language - and produce inspectable, production-grade visual data workflows on Databricks, Snowflake, and BigQuery. The agents do not just suggest code. They build pipelines end-to-end, generate tests, write documentation, and flag errors with proposed fixes.
For Marwaha, this is the product he was describing in abstract when he left Wipro. The tools that existed in 2019 required data engineers, and there were never enough of them. The tools in 2026 require someone who understands the business question. That is a much larger population. And it is the population he spent twenty years selling to.