Unlocking the Future of Healthcare AI: Dr. Jason Moore, Chair of Computational Medicine at Cedars-Sinai
- 1h 9m
- Transcript
Summary
Written by a model from the transcript; not yet reviewed.
This episode explores the career and technical insights of Dr. Jason Moore, chair of the Department of Computational Medicine at Cedars-Sinai. The discussion details his journey from early programming on vintage Atari computers to pioneering contributions in bioinformatics, automated machine learning with T-POT, and multi-agent AI systems for biomedical research. The conversation covers the balance between open-weight models, computational efficiency, and the strict safety and transparency demands required when deploying artificial intelligence in clinical hospital environments.
- Learn how a computer works from the hardware level by programming on vintage computers using assembly language. 8:26
- Develop custom open-source tools and agentic platforms rather than relying entirely on commercial solutions to maintain transparency and trust in healthcare. 34:13
- Utilize knowledge graphs to guide automated machine learning algorithms away from known disease pathways to uncover novel therapeutic drug repurposing opportunities. 49:53
- Build technical skill sets in agentic AI because autonomous agent teams can drastically accelerate complex data analysis and research workflows. 1:03:43
- Continuously question artificial intelligence technologies regarding bias, fairness, safety, privacy, and security before deploying them in society or clinical care. 1:05:19
Chapters
- 0:08Podcast Intro and Host Greeting
The host welcomes Dr. Jason Moore and sets the stage for an in-depth interview on computational medicine.
- 4:33Leadership at Dartmouth and Penn
The interview covers his roles at Dartmouth and Penn, founding departments, PhD programs, and advancing AI and informatics initiatives.
- 6:21Founding the Cedars‑Sinai Computational Medicine Department
Dr. Moore describes establishing a new department, a PhD program, and the hospital’s AI council to integrate AI into patient care.
- 11:21Transition to Faculty and Infrastructure Building
Dr. Moore explains his rapid ascent at Vanderbilt, building bioinformatics cores, HPC resources, and pioneering machine-learning tools like MDR.
- 31:44Knowledge‑Based Agentic AI Development
He details the creation of knowledge graphs, AI agents (Cragan, Escargo), and the ECO XAI platform for autonomous biomedical discovery.
- 42:55Open‑Source AI Platforms and Sustainability
The discussion turns to open‑source models, local deployment, and the ASARI sandbox for agent engineering and experimentation.
- 55:54Professional Recognitions and Future Directions
He reflects on fellowships, interdisciplinary expertise, and future plans in quantum computing, AI, and regenerative medicine.
- 57:02Quantum Investment Landscape
Discussion of the massive capital flow into quantum computing and its potential impact on healthcare.
- 58:41Quantum’s Role in Gene Interaction Research
Exploration of how quantum computing could enable large‑scale enumeration of gene interactions.
- 1:03:11Agentic AI’s Rapid Advancement
Overview of the speaker’s experience with agentic AI and its transformative potential in research.
- 1:03:57Recommendations for AI Adoption
Practical advice for listeners to learn and build agentic AI systems.
- 1:05:36Ethical and Security Concerns in AI
Emphasis on the need to ask hard questions about fairness, safety, and privacy in AI deployments.
Highlights
- 3:24
Atari and Assembly Language Beginningsstory
Learning programming on an Atari 800 in seventh grade sparked a lifelong fascination with computing. Mastering assembly language for machines without GPUs taught deep control over CPU registers and memory management.
- 11:58
Choosing Biology and Complex Systemsstory
Realizing a greater passion for analyzing biological results than performing bench experiments led to graduate studies at the University of Michigan. Focusing on complex systems and nonlinear dynamics in genetics laid the groundwork for a career in computational medicine.
- 14:27
Perfect Timing in Bioinformaticsinsight
Entering academia right as DNA microarrays emerged provided a massive advantage because very few researchers knew how to analyze high-dimensional genomic data. This alignment of training and timing allowed early research programs to scale rapidly.
- 23:29
Automating Machine Learning with T-POTstrategy
Using genetic programming to evolve machine learning pipelines resulted in T-POT, putting the lab two years ahead of major tech companies in automated machine learning software development. The open-source project aimed to democratize machine learning for non-experts.
- 33:19
Overcoming ChatGPT Knowledge Limitationsinsight
Standard large language models failed completely when tested on synthesized biomedical knowledge graphs because they were never trained on integrated domain databases. Specialized retrieval-augmented generation and smart AI agents were required to bridge this gap.
- 45:17
Prioritizing Internal Systems Over Commercial Toolsstrategy
Deploying artificial intelligence inside a hospital requires complete provenance, explainability, and trust for clinicians and IT teams. Building open-source agentic platforms internally ensures full control and transparency over commercial alternatives.
- 49:53
Autonomous Drug Repurposing for Alzheimer'sdata
By instructing agentic AI to look away from known Alzheimer's disease pathways within knowledge graphs, autonomous systems successfully generated hypotheses and identified an existing HIV drug with potential anti-inflammatory effects for Alzheimer's treatment.
Quotes
“I had a very inspirational high school biology teacher who really took the time to teach us college level biology, microbiology in high school.”
Unnamed speaker · 4:04 “And I just was absolutely utterly fascinated with the work he was doing.”
Unnamed speaker · 12:43 “We built the Alzheimer knowledge base, which is publicly available.”
Unnamed speaker · 49:14 “I firmly believe that because I've seen it myself. I've seen the tools we've been able to develop. I've seen how we can take work that used to take us months or a year, and agents can do it in a couple hours or a couple days.”
Unnamed speaker · 1:03:25 “I think it's worth your time and effort to become familiar with Agentic AI, how it works, and to explore tools for building agents. I think it's worth learning about. It's worth developing that skill set, and you'll never regret it.”
Unnamed speaker · 1:03:57 “And once you learn how to engineer a team of agents to do autonomous work for you, it's incredibly empowering. It's a, it really is a superpower.”
Unnamed speaker · 1:03:43 “I think, I think it's our responsibility as humans living in a society to question everything about AI. And a lot of people are doing that.”
Unnamed speaker · 1:05:19
Sponsored segments
- brand called You 1:08:38–1:09:07
About this episode
Discover the pioneering journey of Dr. Jason Moore, Chair of the Department of Computational Medicine at Cedars-Sinai Medical Center, in this episode of The Brand Called You (TBCY). Explore how Dr. Moore has advanced the fields of complex genetics, open-source AI, and biomedical research while contributing to innovative research programs and interdisciplinary education. He shares insights into agentic AI, biomedical data analytics, knowledge-grounded AI, and the democratization of machine learning in healthcare. From his early passion for biology and computing with vintage computers to his work on high-performance computing infrastructure and machine learning at Vanderbilt, Dartmouth, and the University of Pennsylvania, Dr. Moore discusses the development of influential tools such as TPOT and EcoXAI. He also explains why transparency, open-source platforms, and responsible AI development are essential in medicine. Discover how agentic AI can support drug-repurposing research for Alzheimer’s disease, learn about the development of a hospital-embedded Health AI PhD program, and explore Dr. Moore’s views on the future of quantum computing in biomedical research. Dr. Moore concludes with valuable advice for professionals and students interested in agentic AI, emphasizing the importance of developing practical skills while questioning AI’s implications for privacy, fairness, safety, and society. Explore the YouTube chapters below to jump directly to the topics that interest you most.
Transcript
10,933 words · 3 speakers
Topics
- AI
- quantum
- Agentic AI
- T-Pot
- NVIDIA
- XAI
- Eco
- Microsoft
- ECO XAI
Page updated · transcript and summary are machine-generated