
Sieve
Sieve is a YC Winter 2022 (W22) company. Based in San Francisco, CA, USA. Team of 18.
About Sieve
Summary
Sieve builds curated multimodal training datasets, covering video, audio, image, and interaction data, for frontier AI labs and large enterprises. It sources, filters, indexes, annotates, and delivers training-ready data at petabyte scale, replacing the ad-hoc data collection and labeling work that AI research teams would otherwise do in-house. The product targets the bottleneck of high-quality video and multimodal data that limits progress in generative AI, robotics, and AR/VR modeling.
Business model
Pricing is based on data volume, task complexity, and annotation requirements, structured as a purchase agreement per dataset delivery.
Go-to-market
AI research teams and enterprise ML teams are the buyers; they engage by browsing ready-made datasets or requesting custom data samples, then scope and purchase a dataset package directly with Sieve.
From the Y Combinator listing
Video datasets for frontier AI
Sieve is the only AI research lab exclusively focused on video data. Video already makes up 80% of internet traffic and has become the dominant medium driving creativity, communication, gaming, AR/VR, and robotics. Unlocking the ability to truly model video is the key to breakthroughs across all of these domains but progress has been bottlenecked by one thing: high-quality training data. That’s where Sieve comes in. We bring together exabyte-scale video infrastructure, novel video understanding techniques, and dozens of diverse data sources to create datasets that push the frontier of video modeling. This unique combination allows us to deliver data with unmatched precision, quality, and speed which has earned the trust of frontier AI labs, Fortune 100 companies, and fast-growing generative AI startups.
Facts
- B2B / B2B -> Infrastructure
- San Francisco, CA, USA
- 18
- Active
- Growth
Founders (2)
Score every batch against your own thesis
GPAgent's sourcing agent reads every company in a batch, scores each one against your fund's thesis in a private workspace, and lines up the founders worth meeting before Demo Day. You decide who to contact. Your scores, notes, and pipeline never appear on these pages.
How Accelerator intelligence works