# SuperAnnotate

> SuperAnnotate is a San Francisco-based AI data platform that combines annotation and evaluation software with managed expert teams. Founded by researchers working on image segmentation, it now helps enterprise and frontier AI teams curate datasets, run quality reviews, and evaluate models and agents across text, images, video, and audio.

- **Founded:** 2018
- **Headquarters:** San Francisco, United States
- **Founders:** Vahan Petrosyan (CEO and co-founder; formerly CTO), Tigran Petrosyan (Co-founder; formerly CEO), Davit Badalyan (CTO and co-founder), Jason Liang (CBO and co-founder)
- **Products:** Multimodal AI data platform, GenAI and model evaluation workflows, Managed expert workforce, Orchestrate and Python SDK, RL environments and agent trajectories
- **Notable:** Named 2025 Databricks Customer Impact Partner of the Year., Flo Health reported 10x evaluation throughput and 3–5 day iteration cycles in its published case study., Superhuman reported roughly 20 engineering hours saved per week during its evaluation-tool migration.

## Products & services

- **Multimodal AI data platform** — Configurable annotation and evaluation interfaces for text, image, video, audio, LiDAR and agent data, with data curation, project management and quality controls.
- **GenAI and model evaluation workflows** — Rubrics, human review, model-assisted judging, preference data and quality gates for LLMs and agents.
- **Managed expert workforce** — Vetted domain specialists and managed teams for annotation, evaluation and benchmark construction.
- **Orchestrate and Python SDK** — Workflow automation, triggered Python actions and API integrations for moving data and review results through AI pipelines.
- **RL environments and agent trajectories** — Tools and expert services for building tasks, rubrics, rewards and reviews for agent behavior.

## Achievements

- Named 2025 Databricks Customer Impact Partner of the Year.
- Flo Health reported 10x evaluation throughput and 3–5 day iteration cycles in its published case study.
- Superhuman reported roughly 20 engineering hours saved per week during its evaluation-tool migration.
- Wizard reported a 75% cut in weekly annotation costs while maintaining 96% residual accuracy in its case study.

## Latest updates

- **2026-09** — Published Superhuman case study on replacing in-house AI evaluation tooling.
- **2026-03** — Published Wizard and NVIDIA Nemotron case study reporting lower evaluation costs.
- **2026-02** — Published Databricks OfficeQA benchmark collaboration.
- **2025-07** — Announced a $13.5 million Series B extension from Dell Technologies Capital.

## Links

- Website: https://www.superannotate.com/
- LinkedIn: https://www.linkedin.com/company/superannotate
- Twitter/X: https://twitter.com/superannotate
- GitHub: https://github.com/superannotateai
- Facebook: https://www.facebook.com/superannotate

---

Profile page: https://yespress.io/superannotate
Published by YesPress — https://yespress.io
Last updated: 2026-09-29