Mark Bissell
Research and engineering for life sciences and healthcare
San Francisco, California
Joined June 2026
Summary
Mark Bissell is deeply involved in the field of AI interpretability, currently conducting applied research at Goodfire AI. His work focuses on understanding and designing advanced AI systems, including the use of agents for scientific discovery and experimentation. He actively shares insights and research from Goodfire AI, particularly concerning LLMs and their practical applications. markmbissell+2
With a strong background in software engineering and technical leadership, Mark has a proven ability to operationalize complex AI systems for real-world use cases, particularly in the healthcare sector. During his time at Palantir Technologies, he developed LLM orchestration engines and proprietary workflows that led to pending patents related to k-anonymization, showcasing his innovation in privacy-preserving algorithms. markmbissell+1
Mark has significant experience in large-scale data platforms and public health initiatives. As a Forward Deployed Software Engineer Intern at Palantir, he contributed to critical projects like the Tiberius platform for COVID vaccine distribution and the National COVID Cohort Collaborative (N3C) data platform, and co-authored academic papers on clinical data networks. markmbissell+2
Early in his career, Mark demonstrated a keen interest in entrepreneurship, venture capital, and addressing climate change. His fellowship at Prime Coalition involved analyzing philanthropic funding, developing software tools like CRANE to assess greenhouse gas emissions, and researching blended finance, highlighting his commitment to impactful innovation. williams+1
Mark's academic journey includes a Computer Science and Economics degree from Williams College, where he was a Valedictorian and received the Ward Prize for Best Project in Computer Science, and an integrated undergraduate program involving three years at Williams and one at the University of Oxford with the Williams-Exeter Programme. His academic achievements underscore a strong analytical and technical foundation. markmbissell+2
Work
Education
Writing
You and Your Research Agent: Lessons From Using Agents for Interpretability Research
October 2, 2025Authored a blog post on the Goodfire research blog about lessons learned from using agents for interpretability research.
Synergies between centralized and federated approaches to distributed networks for clinical data
January 1, 2022Co-authored a paper on synergies between centralized and federated approaches to distributed networks for clinical data.
Tiberius
Involved in building Tiberius, the platform powering COVID vaccine distribution in the U.S.
k-anonymization patent
Listed on patents for an implementation of k-anonymization, a privacy-preserving algorithm.