# Vik Paruchuri

> Vik Paruchuri is the founder and CEO of Datalab, an AI startup building small, efficient foundation models for document intelligence. A self-taught ML engineer who majored in American History, he previously founded Dataquest - an online learning platform that taught data skills to over 1 million students. His open-source projects (Marker, Surya, Chandra OCR) have earned thousands of GitHub stars and benchmark-topping accuracy scores. He publishes 'The Vik Letter' newsletter covering semiconductors and tech.

- **Role:** Founder & CEO at Datalab
- **Organizations:** Datalab, Dataquest, Answer.ai, Actigram Labs, VCP Analytics
- **Nationality:** American
- **Education:** Bachelor of Arts in American History, University of Maryland - College Park
- **Known for:** Founded Dataquest - taught AI/data skills to 1M+ students over 8 years, Built Marker (PDF-to-markdown converter) - 10x faster than Nougat, #1 on Hacker News with 700 votes in 72 hours, 3,400+ GitHub stars in 3 days, Built Surya OCR - supports 90+ languages, outperforms Tesseract across almost all languages

## Career timeline

- **2004-2009** — BA in American History, University of Maryland (self-described as barely graduating, failed many classes)
- **2006-2009** — Operations Supervisor at UPS
- **2009** — Logistics Supervisor at PepsiCo
- **2010-2011** — Foreign Service Officer, U.S. Department of State (held Active Top Secret Security Clearance)
- **2011-2014** — Data Science Consultant & ML Engineer, VCP Analytics and edX (early team)
- **2012** — Self-taught machine learning and Python via Kaggle competitions
- **2014** — Co-founded Actigram Labs
- **2015-2023** — Founded and ran Dataquest for 8 years, growing to 1M+ students
- **2023-2024** — Learned deep learning from scratch; built open-source OCR tools (Marker, Surya) that went viral
- **2024** — Founded Datalab, raised $3.5M seed round from Pebblebed (OpenAI/FAIR founders)
- **2024** — Released Chandra OCR achieving state-of-the-art benchmark performance (85.9%)
- **2024-present** — Growing Datalab with customers including Anthropic; publishing The Vik Letter newsletter

## Achievements

- Founded Dataquest - taught AI/data skills to 1M+ students over 8 years
- Built Marker (PDF-to-markdown converter) - 10x faster than Nougat, #1 on Hacker News with 700 votes in 72 hours, 3,400+ GitHub stars in 3 days
- Built Surya OCR - supports 90+ languages, outperforms Tesseract across almost all languages
- Built Chandra OCR - 85.9% accuracy on olmocr benchmark (state-of-the-art open-source), 4B parameters
- Raised $3.5M seed round for Datalab from Pebblebed (founded by OpenAI/FAIR founding members)
- Secured Anthropic as a customer for Datalab document intelligence API
- Won multiple Kaggle competitions in automated essay scoring, bond trading, and stock trading
- Built 2 open-source libraries with 5k+ GitHub stars each
- Attracted collaboration offer from Answer.ai (Jeremy Howard's AI lab) based on open-source work
- Contributed 901 commits to Open edX platform
- 89 GitHub repositories, 4,100+ followers on GitHub

## Latest updates

- **2024-10** — Released Chandra OCR achieving 85.9% on olmocr benchmark, trending on HuggingFace with thousands of users
- **2024-06** — Founded Datalab with co-founder Sandy Kwon; raised $3.5M seed from Pebblebed
- **2024-05** — Released Marker v2 with image extraction, better table parsing, commercial licensing
- **2024-04** — Surya OCR achieving state-of-the-art results across 90+ languages
- **2024-01** — Publishing The Vik Letter newsletter focused on semiconductors and tech

## Links

- Website: https://www.vikas.sh/
- LinkedIn: https://www.linkedin.com/in/vikparuchuri
- Twitter/X: https://x.com/VikParuchuri
- GitHub: https://github.com/vikparuchuri
- YouTube: https://www.youtube.com/channel/UCoOd7ubhgUAzL2bK0Ejr6tA
- Instagram: https://www.instagram.com/vikparuchuri/

---

Profile page: https://yespress.io/vik-paruchuri
Published by YesPress — https://yespress.io
Last updated: 2026-04-20