Our Story
A Decade at the Top, Then We Saw an Opportunity
Harvard MBAs with technical backgrounds, we spent the last decade at tier-one investment firms—places where LPs deploy billions across public and private markets. We built quantamental models. We learned how institutional money actually works.
And then we saw something: technology has transformed every other industry, but venture capital hasn't evolved its core processes. We had the skills to change that.
Venture Exploded. The Toolkit Didn't.
Venture assets under management increased 5x over the last decade. The number of firms? Up roughly 25x. Yet the fundamental process remains constrained by human bandwidth and inherent cognitive limits:
- Relationship-driven sourcing (relationships are real, but they're limited)
- Pattern matching on proven success profiles (effective, but inherently limiting)
- Betting on founder quality through time-constrained diligence
This approach has created enormous value. But as capital spreads across more firms, the advantages of differentiation narrow. Most firms compete on the same dimensions.
Meanwhile, vast amounts of data exist that could inform better decisions. Every founder leaves digital footprints—GitHub commits, hiring patterns, shipping velocity. Every company broadcasts market signals—revenue, retention, growth trends. Customer sentiment lives on Reddit and Twitter. Product momentum shows up in YouTube comments and social media activity. But that data requires both systematic data science to analyze the structured signals and LLM interpretation to extract signal from qualitative data. This is what traditional investors don't have time to do.
Combining Investment Judgment with Data-Driven Signal Analysis
That's the real insight. We don't replace investor judgment—we scale it. Great investors have always relied on intuition shaped by deep experience. But intuition works best when tested against data. If your gut says "inspirational founder with product-market-fit," let's verify that against actual signals and comparable companies.
We built tools to combine both. We can screen companies at scale with efficiency that frees up human judgment for the decisions that matter. We can do initial evaluation at 100x the velocity with 1/10 the overhead of a traditional early stage VC. For complex diligence decisions, our tools surface the signal that matters, faster.
This isn't a replacement for venture investing. It's an evolution: investor judgment supercharged by data.
What We've Built
Founder identification that works
We identify exceptional founders through their execution. Not credentials. GitHub commits, hiring patterns, and shipping velocity tell us who actually builds—and we find them before they're founders.
Proof, not intuition
We identify companies that have achieved product-market fit and are accelerating—but we see it before traditional financial data shows it. While others look at revenue and retention metrics, we analyze customer reviews, internet chatter, hiring patterns, and digital footprints to detect whether a company is plateauing, accelerating, or declining. We see momentum shifts months before they appear in financial statements. We don't guess based on a pitch deck or founder conviction. We measure what's actually happening in the market.
Data replaces guesswork
Where traditional investors rely on interviews and pattern matching to assess traction, we look at actual market signals. Revenue, retention curves, user growth, hiring pace. The truth is in the data, not in a founder's conviction.
Why Now
The infrastructure exists. Founders leave digital footprints everywhere—GitHub commits, hiring patterns, customer reviews, social signals, API usage, job postings. This data was invisible a decade ago. Now it's abundant.
The toolkit exists. We gather this data systematically, run rigorous data science analysis to identify patterns, and use modern language models to reason about unstructured signals at scale. What requires days of human analysis, we can process in seconds—across thousands of companies.
Systematic market signal analysis—combining data science on structured signals with LLM interpretation of qualitative data—makes it possible to reduce bias while improving decision quality. This wasn't possible before. We can measure what's actually happening, not guess.
That changes the outcome. Better founders actually get funded. Better companies actually get backed. And investment decisions become more rigorous and data-informed.