David Talby

Career

Before leading John Snow Labs, Talby built and ran web-scale software platforms and teams at startups, for Microsoft's Bing, and to scale Amazon's financial systems. At John Snow Labs, Talby is chief executive officer, helping healthcare and life science companies put AI to good use.¹

Open-source work

Talby is credited as the creator of Spark NLP, an Apache-2.0 natural language processing library built on Apache Spark ML. The library's citation paper, "Spark NLP - Natural language understanding at scale", was written by Veysel Kocaman and Talby and published in the journal Software Impacts in 2021, after a 2021 arXiv preprint.¹²³

Research

Talby's papers, mostly with John Snow Labs colleagues such as Veysel Kocaman, apply NLP to clinical and biomedical text. They include "Biomedical Named Entity Recognition at Scale" (CADL 2020), work on automated de-identification of large clinical text datasets presented at ML4H 2023, and a 2025 study of whether zero-shot commercial APIs meet regulatory standards for clinical de-identification.⁴⁵⁶

Recent work turns to evaluating medical large language models, including a red-teaming framework for safety, robustness and fairness prepared for the Text2Story 2026 workshop. Talby also writes articles and benchmarks on the John Snow Labs blog, such as 2026 comparisons of clinical de-identification tools.⁷¹

Work and contributions

• Spark NLP Open-source code

  Credited as the creator of Spark NLP on John Snow Labs' author page; the Apache-2.0 repository's citation entry lists Veysel Kocaman and David Talby as authors.¹²

Timeline

1. Nov 12, 2020Co-authored "Biomedical Named Entity Recognition at Scale" with Veysel Kocaman.⁴
2. 2021The Spark NLP citation paper was published in the journal Software Impacts.²
3. Jan 26, 2021Released the Spark NLP paper "Natural Language Understanding at Scale" on arXiv with Veysel Kocaman.³
4. Dec 13, 2023Co-authored a paper on automated de-identification of large real-world clinical text datasets for ML4H 2023.⁵
5. Mar 21, 2025Co-authored a study of zero-shot commercial APIs for clinical text de-identification, accepted at Text2Story 2025.⁶
6. Apr 15, 2026Co-authored a multi-domain red-teaming framework for evaluating medical large language models.⁷

Other documented roles

Roles the sources document that are not confirmed as current.

• Built and ran web-scale software platforms and teams for Bing, Microsoft (Bing)¹
• Built and ran software platforms and teams to scale Amazon's financial systems, Amazon¹

Where they work

• John Snow Labs 1 person in the wiki

Related people

• Fabian PedregosaML frameworks and libraries · Language models
• Samy BengioML frameworks and libraries · Language models
• Lewis TunstallLanguage models · ML frameworks and libraries
• Łukasz KaiserLanguage models · ML frameworks and libraries
• Michael AuliLanguage models · ML frameworks and libraries
• Gabriel SynnaeveLanguage models · ML frameworks and libraries
• Thomas WolfML frameworks and libraries · Language models
• Laurent SartranLanguage models · ML frameworks and libraries
• Joe SpisakML frameworks and libraries · Language models
• Nils ReimersLanguage models · ML frameworks and libraries