Soumyadeb Mitra

Founder and Chief Executive Officer RudderStack

Last Reviewed: September 18, 2026

Executive Summary

Soumyadeb Mitra is the founder and Chief Executive Officer of RudderStack, a customer data platform founded in 2019 on an architectural argument rather than a feature one. Conventional customer data platforms collect a company’s behavioural and customer data into the vendor’s own systems, where it is unified, segmented and piped out to marketing and analytics tools – which means the most valuable asset the company has ends up living somewhere it does not control. RudderStack is warehouse-native: the data lands in the customer’s own Snowflake, BigQuery or Databricks warehouse, and the platform operates on it there. The practical consequences are that the data team keeps ownership and governance, the modelling happens where the rest of the company’s data already is, and switching vendors does not mean losing the asset. The product is built on an open-source foundation, which the company presents explicitly as insurance against lock-in, and carries HIPAA, SOC 2 and GDPR compliance. Mitra came to the problem from machine learning – he was Senior Director of Machine Learning at 8x8 and had previously co-founded the AI marketing analytics startup Mariana.

Career Highlights

• Founded RudderStack in 2019 and has led it as CEO throughout.
• Built the platform warehouse-native, so customer data is collected, unified and activated inside the customer’s own data warehouse rather than a vendor silo.
• Built on an open-source foundation, framed by the company as a structural defence against vendor lock-in.
• Grew the company to roughly 104 employees serving more than 550 customers.
• Achieved HIPAA, SOC 2 and GDPR compliance, reflecting a security and privacy posture aimed at regulated buyers.
• Previously co-founded Mariana, an AI marketing analytics startup.
• Served as Senior Director of Machine Learning at 8x8 immediately before founding RudderStack.
• Earlier engineering experience at KritiKal Solutions.
• Currently steering RudderStack toward an agentic customer data platform, with AI-native workflows for data teams and natural-language self-service for business users.

Professional Journey

Mitra’s career began in engineering, with early work at KritiKal Solutions, and developed into machine learning. He co-founded Mariana, a startup applying artificial intelligence to marketing analytics, and subsequently served as Senior Director of Machine Learning at 8x8, the cloud communications company. Both roles put him on the consuming end of customer data infrastructure: machine learning applied to customer behaviour depends entirely on having clean, complete, governed event data, and the tooling available to get it was – in his assessment – structured against the interests of the people who needed it most.

That assessment is the foundation of RudderStack, which he founded in 2019. The incumbent model routed a company’s customer event stream through a vendor’s infrastructure, where it was stitched into profiles and forwarded to downstream tools. It worked well for marketing teams and badly for data teams, who ended up with their most important dataset held outside the warehouse where every other analysis in the company happened, subject to the vendor’s schema, retention policy and pricing. RudderStack inverted the arrangement: collect the events, but deliver and operate on them inside the customer’s own warehouse.

Two further decisions followed from the same reasoning. The product was built open source, which the company states plainly as a lock-in argument – a customer who can read and run the code is not captive. And compliance was treated as architecture rather than paperwork, with the HIPAA, SOC 2 and GDPR certifications that regulated buyers require, supported by a design in which sensitive data never leaves the customer’s control in the first place.

The company has grown to around 104 people serving more than 550 customers. Its current direction, in its own words, is to build a customer data platform that is agentic through and through – letting data teams work in AI-native workflows while business teams self-serve through natural language rather than filing requests. That is a natural extension for a founder who came from machine learning: the value of putting customer data in the warehouse increases considerably once the systems acting on it can reason rather than merely query.

Education

No education record could be established in this review. RudderStack publishes no founder biography on its own site, and the third-party sources that would normally carry one were not retrievable during this pass. This is an open gap rather than a confirmed absence.