Tesorai's AI Platform Turns Raw Mass Spectrometry Into a Searchable Catalog

The former Verily team has raised $8 million to structure the messy, foundational data that trains biological AI models.

About Tesorai

Published

In a lab, the most valuable data is often the messiest. Raw mass spectrometry files from proteomics experiments contain the foundational signals of biology, but they are notoriously difficult to search, harmonize, and reuse. For the AI models now being trained to discover drugs and biomarkers, this raw data is the essential fuel that too often stays locked away. Tesorai, a San Diego startup founded in 2024, is building the catalog and the search engine to turn those files into a structured, queryable resource [SiliconANGLE, November 2026].

The Wedge in Raw Data

Most bioinformatics tools work on summarized results, the tables and figures that end up in a published paper. Tesorai's differentiation is its focus on the raw experimental files themselves, the DDA and DIA mass spectrometry outputs that are typically archived and forgotten after an experiment concludes [SiliconANGLE, November 2026]. The company's platform is a three-part suite aimed at making this data accessible.

  • Tesorai Search. This is the core engine, a cloud-based tool for processing and searching raw mass-spec files. It is built on a proprietary deep-learning model that the company claims enhances peptide and protein identification without relying on traditional post-processing tools like Percolator [Welcome! | Tesorai, Unknown]. The model has been peer-reviewed and published in the Journal of Molecular Biology [Tesorai | AI Proteomics Platform for Mass Spec Data Analysis, Unknown].
  • Tesorai Chat. This layer sits on top, allowing for conversational analysis of processed datasets. It is designed to handle workflows that would otherwise require dedicated bioinformatics expertise [SiliconANGLE, November 2026].
  • Tesorai Trove. To provide context and validation, the company is curating a catalog of structured, AI-ready public proteomics datasets for cross-study comparison [SiliconANGLE, November 2026].

The initial buyers are biopharmaceutical companies and research institutions using the platform for drug discovery and biomarker validation. A secondary, strategic customer is any company building biological AI models that need high-quality, harmonized training data [Genetic Engineering & Biotechnology News, November 2026].

The Verily-Ready Founding Team

The company's technical credibility stems directly from its three co-founders, who all worked together at Alphabet's life sciences company, Verily.

Role Name Prior Role at Verily
Co-Founder & CEO Melissa Miao Led business initiatives and strategic partnerships [Genetic Engineering & Biotechnology News, November 2026]
Co-Founder & CTO Peter Cimermancic, PhD Led artificial-intelligence teams [Genetic Engineering & Biotechnology News, November 2026]
Co-Founder & CSO Charlie Kim, PhD Launched and led the Molecular Science program [Genetic Engineering & Biotechnology News, November 2026]

This blend of commercial, AI engineering, and deep domain science experience is a clear asset. Kim, in particular, brings a rare perspective as a scientist who also founded and bootstrapped a prior venture, Next Jump, to scale without venture capital [Charlie Kim | HuffPost, Unknown]. The team has extended its bench with several PhD-level researchers and collaborators listed on its peer-reviewed publication [Tesorai Search: Large pretrained model boosts identifications in mass spectrometry proteomics without the need for Percolator | bioRxiv, August 2024].

Funding and Early Traction

In November 2026, Tesorai closed a $5.3 million seed round, bringing its total disclosed funding to $8 million [WFMZ / PR Newswire, November 2026]. The investor syndicate is notably diverse, including Perplexity Fund, IQT, and a mix of venture firms like South Central Ventures and New Stack Ventures [recodex.pro, Unknown]. The company stated it will use the capital to expand work with biopharma and research customers, grow tool partnerships, and extend the platform beyond proteins to other molecular data types [WFMZ / PR Newswire, November 2026].

Tesorai reports it has already signed contracts with biopharmaceutical and research organizations, though it has not publicly named those customers [Genetic Engineering & Biotechnology News, November 2026]. It has cited collaborations with institutions like MD Anderson Cancer Center and New York University [citybiz.co, Unknown]. An early claim of technical validation points to the platform recovering an Alzheimer's disease marker that five prior brain studies had not reported [techfundingnews.com, Unknown].

The Technical Breakdown

Under the hood, Tesorai Search represents a specific architectural bet. The model is a large pretrained neural network that was trained on 21 publicly available datasets, all from Orbitrap mass spectrometers [Back to Basics: Spectrum and Peptide Sequence are Sufficient for Top-tier Mass Spectrometry Proteomics Identification | bioRxiv, August 2024]. By focusing solely on the peptide spectrum and sequence data, the company's research argues it can achieve top-tier identification performance, bypassing the need for the Percolator tool that has been a standard in the field for over a decade [Tesorai Search: Large pretrained model boosts identifications in mass spectrometry proteomics without the need for Percolator | bioRxiv, August 2024].

Operationally, the service runs on cloud-native infrastructure, dynamically scaling with T4 Nvidia GPUs to promise faster processing than conventional, on-premise tools [FAQ page, Unknown]. For customers, this means they can start with individual files. Public pricing lists a Plus plan at $20 per month and a Pro plan at $100 per month, with custom Enterprise plans available [Frequently Asked Questions | Tesorai, Unknown].

Scaling the Data Moat

The long-term bet is that structuring the world's raw proteomics data creates a defensible asset. The more data processed through Tesorai Search, the better the underlying models become at identification, and the more valuable the Trove catalog grows for validation studies. This creates a potential network effect where new customers contribute to and benefit from a continuously improving central resource.

However, the path to scale is paved with technical and commercial friction. The model's training was specific to Orbitrap instruments from Thermo Fisher Scientific, which, while common, is not the entire universe of lab equipment. Generalizing performance across instrument types and experimental conditions will be a continuous engineering challenge. On the commercial side, moving from early research contracts to enterprise-wide deployments at large pharma companies involves sales cycles and integration work that tests any young team. Furthermore, the platform's success hinges on convincing researchers and institutions to upload their raw data, which may involve navigating data sovereignty and intellectual property concerns that are paramount in competitive drug discovery.

For now, Tesorai has assembled the right team, secured credible funding, and shipped a peer-reviewed core technology. Its progress will be measured by the volume and variety of data flowing through its search engine, and by the drug discovery pipelines that eventually cite it as a critical tool.

Sources

  1. [SiliconANGLE, November 2026] Tesorai raises $5.3M to feed starving AI models with biological data and accelerate drug discovery | https://www.siliconangle.com/2026/11/18/tesorai-raises-5-3m-to-feed-starving-ai-models-with-biological-data-and-accelerate-drug-discovery/
  2. [Genetic Engineering & Biotechnology News, November 2026] Former Verily Leaders Build AI Platform to Get More from Proteomics Data | https://www.genengnews.com/topics/artificial-intelligence/former-verily-leaders-build-ai-platform-to-get-more-from-proteomics-data/
  3. [WFMZ / PR Newswire, November 2026] Tesorai Raises $5.3 Million to Put Vast Troves of Untapped Research Data to Work | https://www.wfmz.com/news/pr_newswire/pr_newswire_stocks/tesorai-raises-5-3-million-to-put-vast-troves-of-untapped-research-data-to-work/article_ffef9a27-3563-58a5-9f4e-dacdbb62f137.html
  4. [recodex.pro, Unknown] Tesorai raises $5.3M seed round to unlock raw lab data | https://recodex.pro/tesorai-raises-5-3m-seed-round-to-unlock-raw-lab-data/
  5. [Tesorai | AI Proteomics Platform for Mass Spec Data Analysis, Unknown] Tesorai | AI Proteomics Platform for Mass Spec Data Analysis | https://www.tesorai.com/
  6. [Welcome! | Tesorai, Unknown] Welcome! | Tesorai | https://www.tesorai.com/
  7. [Tesorai Search: Large pretrained model boosts identifications in mass spectrometry proteomics without the need for Percolator | bioRxiv, August 2024] Tesorai Search: Large pretrained model boosts identifications in mass spectrometry proteomics without the need for Percolator | https://www.biorxiv.org/content/10.1101/2024.08.16.608295v1
  8. [Back to Basics: Spectrum and Peptide Sequence are Sufficient for Top-tier Mass Spectrometry Proteomics Identification | bioRxiv, August 2024] Back to Basics: Spectrum and Peptide Sequence are Sufficient for Top-tier Mass Spectrometry Proteomics Identification | https://www.biorxiv.org/content/10.1101/2024.08.16.608295v1
  9. [FAQ page, Unknown] FAQ page | https://www.tesorai.com/
  10. [Frequently Asked Questions | Tesorai, Unknown] Frequently Asked Questions | Tesorai | https://www.tesorai.com/
  11. [citybiz.co, Unknown] Tesorai raises $5.3 Million Seed round to expand AI-powered life sciences research platform | https://www.citybiz.co/article/935676/tesorai-raises-5-3-million-seed-round-to-expand-ai-powered-life-sciences-research-platform/
  12. [techfundingnews.com, Unknown] Tesorai raises $5.3M seed, backed by Perplexity Fund and IQT, to make raw lab data searchable | https://techfundingnews.com/tesorai-raises-5-3m-seed-backed-by-perplexity-fund-and-iqt-to-make-raw-lab-data-searchable/
  13. [Charlie Kim | HuffPost, Unknown] Charlie Kim | HuffPost | https://www.huffpost.com/author/charlie-kim

Read on Startuply.vc