A computer vision model in under an hour. That is the promise EyePop.ai is selling to startups and SMBs that cannot afford to hire a machine learning team. The San Diego-based company, founded in 2023, has raised $2.85 million to back the claim that you can train a custom model to detect, measure, and count objects from your own images and video without writing a line of code [Crunchbase, 2025].
The wedge is time, not technology
The platform offers two paths. A library of pre-built models for common tasks like detecting people or reading text provides a starting point. The core product, however, is the custom training engine. Users upload their own images or video, label the objects they want the AI to recognize, and the system returns a deployable model via API or SDK, a process the company says takes less than 60 minutes [PR Newswire, April 2024].
A founding team built for product-market fit
The credibility of the bet rests partly on the founders. The team is led by three entrepreneurs, including Andy Ballester, who co-founded the crowdfunding giant GoFundMe [PR Newswire, April 2024]. Ballester serves as Chief Product Officer. CEO Brad Chisum previously sold his company Lumedyne Tech to Google for a reported $85 million. CTO Torsten Schulz rounds out the technical leadership [Employbl].
| Founder | Role | Notable Background |
|---|---|---|
| Brad Chisum | CEO | Previously sold Lumedyne Tech to Google |
| Andy Ballester | Co-founder & CPO | Co-founder of GoFundMe |
| Torsten Schulz | Co-founder & CTO | Board Member at Drewag Stadtwerke Dresden GmbH |
Traction and external validation
EyePop.ai is still early, but it has gathered momentum beyond its seed financing. The company demonstrated a Video Intelligence Agent in collaboration with Qualcomm at the Snapdragon Summit in 2025 [eyepop.ai blog]. More concretely, it won the Judges' Choice Award at the ISC West 2026 security conference for its platform in the Video Analytics category [Security Industry Association, 2026].
Where the model could break
The market EyePop.ai is targeting is attractive but fraught with competition and technical constraints. The company's most credible risks are not hypothetical.
- The crowded middle. The space between open-source frameworks and full-stack enterprise AI platforms is getting dense.
- The performance ceiling. A model trained in an hour on a startup's limited dataset may work for a proof of concept but could struggle with the edge cases and accuracy demands of a production environment.
- The scaling question. The economics of supporting thousands of unique, small-batch custom models are different from serving one large, generalized model.
The next twelve months
The $2.85 million seed round, led by Innosphere Fund with participation from Interlock, Spatial Capital, and Keshif Ventures, gives EyePop.ai runway to prove its model [eyepop.ai blog, March 2025]. The capital will likely be deployed to sharpen the product, build out sales channels, and gather more public case studies.