Research Scientist
Roboflow, San Francisco
About the job
Who We Are
Roboflow is building the visual cortex for AGI.
Every day, millions of developers and half of the Fortune 100 use Roboflow to count cells to accelerate cancer research, improve construction site safety, digitize floor plans, preserve coral reef populations, guide drone flight, and much more.
We have built the tools, community, and resources needed to make the world programmable with artificial intelligence. We're supported by great customers and investors, having raised over $100 million from Google Ventures, Craft Ventures, Sam Altman, Lachy Groom, Y Combinator, and more.
Roboflow Labs is our applied research unit. We train our own open models (RF-DETR). We build our own benchmarks grounded in real enterprise use cases (RF100-VL). And we partner with frontier labs to deliver the training data they need to advance the capabilities that drive the most value back to our customers.
Behind all of it is a global network of millions of Roboflow developers who have annotated billions of images and videos to date, as well as a proven enterprise business that is exploding as AI traverses the physical economy.
What You'll Do
As a Research Scientist at Roboflow Labs, you'll sit at the intersection of our benchmarks, our models, and the researchers at frontier labs who buy our data: running failure analyses that show exactly where a model breaks on real-world tasks, turning those findings into dataset and evaluation specifications, and validating that the data we ship actually moves the numbers. You'll publish, extend RF100-VL and the benchmarks that follow it, and help set the research agenda for a team whose work is grounded in millions of real deployments rather than fabricated test sets.