Small AI Teams with Huge Impact — Vik Paruchuri, Datalab
Jul 15, 2025 · 17:36
Vikas Paruchuri, CEO of Datalab, argues that small teams can outperform large ones, sharing how his team of 4 achieved 40k GitHub stars, 7-figure ARR, and 5x revenue growth since January by training state-of-the-art models like surya OCR 3. Drawing from his experience scaling DataQuest to 30 people and then cutting to 7, he explains that layoffs increased productivity due to fewer specialists, less meeting overload, and more senior generalists. He advocates hiring senior generalists who work across the stack, using simple tech (e.g., server-rendered HTML over React), and minimizing bureaucracy with high trust and in-person collaboration. By leveraging AI to handle low-leverage tasks and training models to replace forward-deployed engineers, Datalab maintains tight feedback loops and fast iteration. Paruchuri emphasizes scaling productivity, not headcount, and offers a three-step hiring process: a peer chat, a paid 10-hour project, and a culture fit check, with a 40% hire rate.
