Knowtex Lands VA's $776M Ambient Scribe Contract and Recruits Engineers in San Francisco

Knowtex, a San Francisco clinical AI company founded in 2022, was named one of two prime awardees on the U.S. Department of Veterans Affairs (VA) Ambient Scribe Enterprise contract. Awarded by the VA Technology Acquisition Center, the multiple-award indefinite delivery, indefinite quantity (IDIQ) contract carries a shared maximum ceiling of $775.72 million across a five-year ordering period. The contract expands ambient clinical documentation tools across the largest integrated healthcare system in the United States.

Originally launched out of Y Combinator's S22 batch, Knowtex has expanded its footprint from an initial $15 million VA pilot in October 2025 to deployment across 79 VA medical centers and more than 310 facilities nationwide. Over 7,000 VA clinicians currently use the platform. To support this expanding deployment, Knowtex is hiring a Senior Software Engineer, Android ($180,000–$220,000/year, San Francisco, hybrid) to build and scale mobile applications within VA health facilities while utilizing agentic coding tools throughout the software development lifecycle.

Specialty-Specific Clinical AI Models

Knowtex builds an artificial intelligence platform that converts clinical conversations into complete medical notes, billing codes, order entries, and clinical intelligence. Rather than relying on generic large language models, the company trains specialty-specific models tailored to more than 200 medical specialties, including oncology, orthopedics, cardiology, and primary care. The system integrates into electronic health record (EHR) platforms such as Epic, Oracle, Athena, Flatiron, and the VA's legacy Computerized Patient Record System (CPRS).

In September 2026, Knowtex launched a healthcare AI research lab and introduced KnowBench, a clinical AI benchmark published on arXiv that evaluates Effort Reduction—the percentage of AI-generated documentation accepted without edits by a clinician. Knowtex reported an effort reduction score of 97.99% on the benchmark.

To develop its underlying model architecture, the company is recruiting an ML Engineer, LLM's (San Francisco, onsite) to optimize open-weight large language models on proprietary clinical datasets. The role focuses on improving inference quality, latency, control, and computational costs across structured reasoning pipelines.

Founder Origins in Medical Scribing

Knowtex was founded by Stanford University AI researchers Caroline Zhang (CEO) and Jocelyn Kang (CTO). Prior to launching the company, Zhang and Kang spent six months working in scrubs as medical scribes—Zhang in oncology clinics and Kang in orthopedic surgery—to observe patient interactions, documentation workflows, and order placement firsthand.

Zhang holds a master's degree in computer science with an AI concentration from Stanford, an undergraduate degree in economics, and a master's degree from Peking University. Her background includes biomedical research at the Salk Institute, investment banking at Morgan Stanley, AI policy work at Schmidt Futures, and legislative policy work in the U.S. Senate. Kang graduated Phi Beta Kappa and Summa Cum Laude from Stanford with a degree in Symbolic Systems and a minor in philosophy. She previously led product and engineering initiatives at startups including Learn to Win.

In an interview on the Techy Surgeon podcast, Zhang described ambient clinical technology as a core operating system for clinicians, explaining that direct voice capture enables real-time synthesis of patient medical histories alongside active clinical protocols. Speaking on the Delta: HealthTech Innovators podcast, Zhang noted that automated documentation surfaces ICD-10, CPT, and Evaluation and Management (E/M) codes directly at the point of care, helping health systems capture administrative revenue while keeping documentation accurate.

Enterprise Expansion Across Federal and Commercial Health Systems

Knowtex entered the VA system after winning the department's 2024 AI Tech Sprint against a field of 150 competitors, as reported by Yahoo Finance and HCI Innovation Group. The initial pilot launched at the Kansas City VA Medical Center in October 2025 before expanding to facilities in Dallas, Miami, Loma Linda, and East Orange. In February 2026, the VA granted national production approval for primary care across 10 Veterans Integrated Service Networks (VISNs).

According to a white paper cited by PR Newswire, participating clinicians at the Kansas City site reported saving one to two hours per day of after-hours documentation time. Data published by the VA indicated patient trust ratings rose to 95% during the initial 90-day evaluation. As noted by ABC15 Arizona, audio recordings are maintained under federal compliance guidelines for up to 30 days before permanent deletion, with clinicians retaining mandatory review and sign-off authority for all medical records.

Beyond federal contracts, Knowtex partners with commercial health organizations and specialty practices. In a video on Caroline Zhang's YouTube channel, oncology network cCARE detailed how integrating Knowtex reduced daily documentation burdens by two hours per clinician. The company also established a technical partnership with 4D EMR, reported by PRWeb, to integrate voice workflows natively into specialty electronic medical records.

To build out user-facing web tools and client-side EHR integrations, Knowtex is hiring a Senior Software Engineer, Web ($180,000–$220,000/year, San Francisco, hybrid). The position focuses on maintaining AI assistant features and application architectures built with TypeScript, React, and Node.js.

Open Engineering Roles at Knowtex

Knowtex operates as a six-person team headquartered in San Francisco. Backed by investors including Gaingels, StartX, HF0, Soma Capital, Pioneer Fund, and Y Combinator, the company is scaling its core engineering group to manage enterprise healthcare deployments.

See all open roles at Knowtex