# Harell Data: Bellevue AI Profile — Product, Footprint & Verified Record

- **Canonical Web Article:** https://tygartmedia.com/harell-data-bellevue-wa-ai-profile/
- **Profile Brief (PDF):** https://tygartmedia.com/wp-content/uploads/2026/11/harell-data-bellevue-wa-ai-profile.pdf
- **Markdown Twin:** https://tygartmedia.com/wp-content/uploads/2026/11/harell-data-bellevue-wa-ai-profile.md
- **Snapshot Infographic:** https://tygartmedia.com/wp-content/uploads/2026/11/harell-data-bellevue-wa-ai-profile-infographic.png
- **Verified Date:** 2026-11-19

> **Sourced Baseline:** As of November 19, 2026, Harell Data Corp. operates Harell Cloud at harelldata.com, connecting AI modelers with proprietary scientific datasets from early partners including Adaptive Biotechnologies and A-Alpha Bio. The Bellevue startup emerged from stealth in August 2026 with $15M in funding from Fuse, Cercano Management, and other backers, and joined the Fall 2026 GeekWire 200.

## Verified Company Record

| Field | Verified Record (As of Nov 19, 2026) |
| :--- | :--- |
| **Company** | Harell Data (Harell Data Corp.) |
| **Official Website** | https://harelldata.com/ |
| **Washington HQ** | Bellevue, WA (also maintains a Palo Alto, CA office) |
| **Other WA Footprint** | Initial data partners include Seattle-based A-Alpha Bio and Seattle-based Adaptive Biotechnologies |
| **Founded** | 2026 (GeekWire 200 Fall 2026; emerged from stealth August 17, 2026) |
| **Founders / Leadership** | Harlan Robins, PhD (Founder & President), Rakesh Nair (CTO), Saray Covey (Head of Operations), Analise Polsky (Head of Sales) |
| **Stage** | Early / Seed |
| **Announced Funding** | $15M (Fuse, Cercano Management, and others; GeekWire, Aug 17, 2026) |
| **Team Size** | 11 reported at August 2026 launch (Bellevue and Palo Alto); 9 listed on Fall 2026 GeekWire 200 |
| **Primary Ecosystem Ties** | GeekWire 200 (Fall 2026); Fuse; Cercano Management; Adaptive Biotechnologies; A-Alpha Bio |

## What Harell Data Builds

Harell Data Corp. builds **Harell Cloud**, a secure cloud infrastructure and data-marketplace platform that connects AI model builders with curated proprietary scientific datasets without requiring data owners to hand over raw files. As [GeekWire reported on August 17, 2026](https://www.geekwire.com/2026/adaptive-biotech-co-founder-raises-15m-for-new-startup-to-rethink-how-ai-trains-on-science/), the company was created to break open scientific data silos: research groups spend millions generating high-value experimental datasets but have lacked a secure commercial mechanism to monetize them for AI training.

On [harelldata.com](https://harelldata.com/), Harell Cloud presents three core layers—**Data**, **Models**, and **Inference**—deployed on [CoreWeave](https://www.coreweave.com/) GPU infrastructure (NVIDIA A100 and Hopper GPUs) so machine learning teams train and fine-tune models in place inside a protected environment. Instead of waiting years for speculative downstream drug royalties, data creators earn a direct share of the compute revenue generated during training runs. While Harell Data is starting in computational medicine and immunology, founder Harlan Robins told GeekWire the architecture is designed to extend across scientific domains including materials science and imaging.

- **Secure in-place training (Data):** Hosts curated proprietary datasets alongside private customer data so researchers train and fine-tune AI models without raw data leaving the cloud environment or negotiating bespoke licensing.
- **Benchmarked model ecosystem (Models):** Provides access to curated scientific models paired with real-time benchmarking and leaderboards designed to recognize and reward contributors.
- **Immunology inference workflows (Inference):** Highlights antibody binding-site and relative binding-strength prediction alongside T-cell receptor (TCR) physiological binding-site and activation-threshold prediction.

## Washington Footprint and Leadership

Harell Data is headquartered in **Bellevue, Washington**, with a split team operating between Bellevue and Palo Alto, California, as reported by [GeekWire on August 17, 2026](https://www.geekwire.com/2026/adaptive-biotech-co-founder-raises-15m-for-new-startup-to-rethink-how-ai-trains-on-science/). Its founding ties run directly through Seattle's biotechnology cluster: founder **Harlan Robins, PhD** co-founded Seattle-based Adaptive Biotechnologies, served as its Chief Scientific Officer building large-scale immune-system datasets, and continues consulting on scientific strategy there. Robins told GeekWire the company's name came from his 8-year-old son Ellis, combining "Harlan" and "Ellis."

Alongside Robins as Founder and President, GeekWire's launch report and company materials on [harelldata.com/about-us/](https://harelldata.com/about-us/) identify an initial executive team including **Rakesh Nair** (Chief Technology Officer), **Saray Covey** (Head of Operations), and **Analise Polsky** (Head of Sales). Headcount varies slightly across two dated primary snapshots: GeekWire reported an 11-person team split between Bellevue and Palo Alto at its August 17, 2026 launch, whereas the Fall 2026 [GeekWire 200](https://www.geekwire.com/geekwire-200/) index listed 9 employees. The active website is harelldata.com, while harelldata.ai holds a "Launching Soon" landing page.

## Products, Users, and Public Traction

Harell Data serves two sides of computational science: life-sciences and research organizations holding proprietary experimental datasets, and AI research teams seeking curated data and GPU compute to train biological foundation and prediction models. Early access for Harell Cloud launched alongside the August 2026 announcement with initial proprietary datasets contributed by Seattle-based **A-Alpha Bio** and **Adaptive Biotechnologies**, per [GeekWire](https://www.geekwire.com/2026/adaptive-biotech-co-founder-raises-15m-for-new-startup-to-rethink-how-ai-trains-on-science/).

Active scientific community usage is also documented on [Harell Data's LinkedIn page](https://www.linkedin.com/company/harell-data/), which details proprietary TCR-peptide datasets provided by Adaptive Biotechnologies Corp. on Harell Data for the **CASP-DREAM TAP Challenge** (organized with Conscience, DREAM Challenges, and Sage Bionetworks)—including more than 1,000,000 unique fully paired alpha/beta TCRs tested against over 1,500 peptides for HLA-A*02:01 and more than 27,000 TCRs tested against over 380 peptides for HLA-B*27:05. Harell Data also partners with [CoreWeave](https://www.coreweave.com/) for NVIDIA GPU cloud capacity. Revenue figures and total paying customer counts are not publicly disclosed.

## Funding, Stage, and Place in Washington's AI Landscape

Harell Data emerged from stealth on August 17, 2026 with **$15 million** in early-stage funding backed by Bellevue-based **Fuse**, **Cercano Management**, and additional unnamed investors, according to [GeekWire](https://www.geekwire.com/2026/adaptive-biotech-co-founder-raises-15m-for-new-startup-to-rethink-how-ai-trains-on-science/). Following its launch, the company entered the Fall 2026 [GeekWire 200](https://www.geekwire.com/geekwire-200/) startup index (listing Bellevue HQ, $15M raised, and 2026 as its founding year). No subsequent financing round or acquisition has been announced through November 2026.

Within Washington's biotech and scientific AI cluster, Harell Data targets the underlying data-access and incentive bottleneck separating wet-lab data generators from AI model developers. While Seattle peers build autonomous experimental agents, live-cell analysis engines, and oncology drug-discovery pipelines, Harell Data provides neutral cloud infrastructure and compute-revenue sharing aimed at making proprietary scientific datasets usable for model training without sacrificing IP control.

## Related Washington AI Profiles & Tygart Media Context

- [Potato — Seattle AI Profile](https://tygartmedia.com/potato-seattle-wa-ai-profile/)
- [Clockwork Bio — Seattle AI Profile](https://tygartmedia.com/clockwork-bio-seattle-wa-ai-profile/)
- [Polybiomics — Seattle AI Profile](https://tygartmedia.com/polybiomics-seattle-wa-ai-profile/)
- [Why AI Compute Needs a Warranty, Not Another Watermark](https://tygartmedia.com/certified-compute-warranty-not-watermark/)

## Verified Primary Sources

- [Harell Data Official Website](https://harelldata.com/)
- [Harell Data About Us](https://harelldata.com/about-us/)
- [GeekWire — Adaptive Biotech Co-Founder Raises $15M for Harell Data (Aug 17, 2026)](https://www.geekwire.com/2026/adaptive-biotech-co-founder-raises-15m-for-new-startup-to-rethink-how-ai-trains-on-science/)
- [GeekWire 200 Startup Index](https://www.geekwire.com/geekwire-200/)
- [Harell Data LinkedIn](https://www.linkedin.com/company/harell-data/)
- [CoreWeave Official Website](https://www.coreweave.com/)

## Frequently Asked Questions

### How does Harell Data compensate scientific data providers?

According to GeekWire's August 2026 launch report, data owners host proprietary datasets inside Harell Cloud so AI models train in place without raw data leaving the platform, and providers earn a direct share of compute revenue generated during training runs rather than waiting on downstream drug royalties.

### Who founded Harell Data and where is it headquartered?

Harell Data was founded by Harlan Robins, PhD, co-founder and former Chief Scientific Officer of Seattle-based Adaptive Biotechnologies. The company is headquartered in Bellevue, Washington, with team members split between Bellevue and Palo Alto, California.

### Which organizations have provided initial datasets on Harell Cloud?

GeekWire reported at launch that early access included initial proprietary datasets from Seattle-based A-Alpha Bio and Adaptive Biotechnologies. Harell Data's LinkedIn also documents Adaptive Biotechnologies' proprietary TCR-peptide datasets hosted for the CASP-DREAM TAP Challenge.
