Shraman Kar

Shraman and his brother Shreyas grew up shipping side projects together, graduating from Arduino kits and hackathons into research at the Stanford AI Lab before either had finished their first year on campus.

Inside the lab they saw a gap: models like Sora and VEO 3 were great at cinematic, meme-friendly clips but bad at explaining a 10‑step onboarding or turning dense training docs into clear, structured video, so they left Stanford to build Golpo AI around that practical use case.

1. They picked a “boring but paid” wedge: instead of chasing entertainment, Golpo turns documents and prompts into animated explainers for education and corporate training, so every demo maps directly to an existing budget line in school districts and enterprises.

2. They framed the product as infrastructure, not novelty, explicitly contrasting Golpo with flashy generative video tools; that narrative made it easy for investors and customers to justify an oversubscribed $4.1M seed as a bet on scalable communication, not speculative creativity.

3. They leaned on institutional credibility and focus: Stanford AI Lab provenance, a clear story of dropping out to chase one glaring gap in AI video, and early pilots with school districts and enterprises created a tight loop where each new proof point de-risked the next conversation.