Bulk outbound got penalized
Relevance is now a deliverability requirement. Spray-and-pray lists burn domains before they create pipeline.
About Qumin
We research before the list exists.
Qumin is building the discovery layer for modern GTM: a product that turns a specific market definition into source-backed accounts, personas, and buyer evidence.
Job boardsNewsGoogle MapsShopifyReviewsICP match
The problem
Static databases were built for bulk prospecting.
Modern outbound starts with highly specific market definitions: teams hiring SDRs, ecommerce brands opening retail stores, companies changing tools, or manufacturers expanding into a new market. Those opportunities are buried in the live web, not waiting as clean rows in a database.
Start in a static database
Filter by fixed fields
Export the matches
Qualify by hand
Reach the same accounts
Why now
The “who fits my ICP?” question is finally answerable.
The public trail exploded
Hiring, funding, reviews, tech-stack changes, launches, and store openings leave more evidence than teams can track by hand.
AI connects the signals
Reasoning agents can read the open web and separate account-fit evidence from noise at software speed.
How it works
Describe your market. Qumin discovers companies others miss.
Evidence trail
Every signal is backed by a verifiable source, so the team can inspect why the account fits.
Persona
Qumin identifies the people connected to each account and waterfall-enriches the data needed to evaluate fit.
Continuous discovery
Saved agents keep surfacing new accounts and get sharper each time you tell Qumin what worked.
Team
Built by people who have shipped the hard part.
Search, signal processing, knowledge graphs, and evidence synthesis are not slideware here. They are the stack the team has already built at scale.
Founder & CEO
Khushal Bapna
IIT Delhi
- 7 years in deep learning, ML, and AI
- Built search and signals systems at Bloomreach
- Advised outbound teams close enough to live the problem
Co-founder & CTO
Srihari Maruthachalam
IIT Madras
- Shipped production AI models for real-world signal data
- Built safety-critical AI agents and computer vision at Netradyne
- Published BCI research with MIT professors
Founding engineer
Dakshina Murthy
IIT Madras
- Built production knowledge graph and vector search systems at Amazon
- Shipped RAG assistants on Bedrock
- Natural-language querying over real knowledge bases
Founder-market fit: search, signal processing, and knowledge graphs shipped at Bloomreach, Netradyne, and Amazon.