Foresite Labs Appoints Dr. Jorge Reis-Filho to Lead AI-Driven Biomedical Discovery Strategy
核心洞察
Foresite Labs (搜索) has appointed Dr. Jorge Reis-Filho as Chief Strategy Officer and Operating Partner, with a concurrent role as Venture Partner at Foresite Capital (搜索).
Dr. Reis-Filho joins from AstraZeneca, where he served as Chief of AI for Science Innovation, Enterprise AI, leading AI initiatives to transform drug discovery and clinical development.
His appointment aims to translate advances in AI, machine learning, statistical genetics, and data science into new companies and platforms at the convergence of AI and life sciences.
Foresite Labs (搜索) has appointed Dr. Jorge Reis-Filho as Chief Strategy Officer and Operating Partner, with a concurrent role as Venture Partner at Foresite Capital (搜索). Dr. Reis-Filho joins from AstraZeneca, where he served as Chief of AI for Science Innovation, Enterprise AI, leading initiatives to transform drug discovery and clinical development through AI, including the development of foundation models and agentic frameworks. In his new role, he will help translate advances in AI, machine learning, statistical genetics, data science, and biology into differentiated technologies, platforms, products, and new companies, while providing strategic and operational oversight across venture formation, governance, team building, financing, and growth.
"Jorge brings a rare combination of expertise across drug development, diagnostic and experimental pathology and artificial intelligence. He has spent his career at the forefront of translating advances in computational medicine and AI into real-world impact and validating AI platforms inside one of the world's leading pharmaceutical companies," said Jim Tananbaum, co-founder and CEO of Foresite Labs (搜索). "That combination of scientific depth and operational experience is exactly what it takes to build enduring companies, and we're excited to have him bring that expertise to Foresite Labs."
A Career Bridging Pathology, Biomarkers, and AI
Prior to joining Foresite Labs (搜索), Dr. Reis-Filho served as Chief AI and Data Scientist for Oncology R&D and Vice President of Cancer Biomarker Development at AstraZeneca, where he led the creation and clinical validation of biomarkers used to match patients with optimal treatments. Before AstraZeneca, he was Chief of Experimental Pathology at Memorial Sloan Kettering Cancer Center, applying genomic technologies, big data analytics, and AI to bridge academic research and commercial innovation.
Dr. Reis-Filho is an expert in molecular pathology, bioinformatics, functional genomics, and AI. His work has been recognized with honors including the CRUK Future Leaders Prize, the BCRF Larry Norton Award, and the William Gerald Award from MSKCC. He earned a joint MD from the University of Porto, Portugal, and the Universidade Federal do Paraná, Brazil; completed histopathology training at the University of Porto; and obtained a PhD in breast cancer molecular pathology from the Institute of Cancer Research and the Royal Marsden Hospital in London.
"My career has been driven by translating scientific insights into biomarkers, technologies and therapeutics that transform outcomes for patients, first in academic pathology and, more recently, through building AI platforms within a global biopharmaceutical company," said Dr. Reis-Filho. "I am immensely excited to bring that experience to Foresite Labs (搜索) and help build the next generation of enduring companies at the convergence of AI and the biomedical sciences."
A Track Record of Incubation Milestones
Dr. Reis-Filho's appointment comes as Foresite Labs (搜索)' incubation platform continues to deliver major milestones. In 2024, the company co-incubated Xaira Therapeutics (搜索) alongside ARCH Venture Partners (搜索), which launched with more than $1 billion in committed capital. That same year, Candid Therapeutics (搜索) launched with $370 million to advance novel T-cell engager therapies for autoimmune disease and was later acquired by UCB (搜索) for $2.2 billion. Alumis, focused on precision immunology, completed its IPO that same year, and most recently, Foresite Labs co-founded Prometheus (搜索), the physical AI company which has since raised more than $12 billion, with ARCH Venture Partners.
From AI Agents to AI Innovators
In a detailed essay shared on LinkedIn, Dr. Reis-Filho outlined his vision for the next inflection point in artificial intelligence: the emergence of "AI innovators," systems capable of originating novel and falsifiable hypotheses, identifying the experiments most likely to distinguish among competing explanations, learning from negative evidence, and generating knowledge that survives independent validation.
"Artificial intelligence is approaching its most important transition: from systems that retrieve, reason and execute within an existing evidence base to systems capable of expanding that evidence base through discovery," Dr. Reis-Filho wrote. He distinguishes this from current capabilities: "agents operate on an existing evidence base, whereas innovators expand it by generating novel hypotheses and the evidence required to resolve questions for which the correct answer is not yet knowable, a capability known as open-endedness."
Dr. Reis-Filho described an exercise in which two teams of AI agents revisited 20 high-stakes biopharma decisions using only the evidence available at each historical decision point. "In some instances, the agents arrived at better decisions than those made at the time; in others, they reproduced the failure because the evidence required to make the correct choice did not yet exist," he wrote. "This exercise illustrated both the remarkable power and the present ceiling of agents."
Why Biology Lags Software and Mathematics
Dr. Reis-Filho argues that software engineering and mathematics moved first toward domain-bounded superhuman capability for structural reasons: their objects can be represented in machine-legible form, candidate solutions can be evaluated rapidly, and evaluators are often objective. He notes that an advanced version of Gemini Deep Think achieved gold-medal performance at the 2025 International Mathematical Olympiad, while AlphaEvolve has improved solutions to open mathematical problems.
"Biology has almost the opposite structure," he wrote. "Its state is only partially observed and, more often than not, imprecisely measured... Experiments may take weeks or years, cost millions of dollars and yield noisy or ambiguous results; an intervention changes the very system a model is attempting to predict, and at the clinical scale the most meaningful evaluator may be a patient outcome observed years later."
He identifies five interlocking requirements for achieving biomedical superintelligence: representing biology at its native resolution, making time a first-class dimension, closing the causal loop with experimentation and falsification, defining progress with fit-for-purpose evaluations and benchmarks, and building the system around the model.
Computational Pathology as a Proof Point
Dr. Reis-Filho highlights concrete examples of AI translating into clinical utility. He describes Quantitative Continuous Scoring (QCS), which generates continuous, single-cell measurements of target expression within specific subcellular compartments, including a normalized membrane-to-cytoplasm ratio that visual inspection cannot reliably determine. In TROPION-Lung01, this feature predicted differential outcomes with datopotamab deruxtecan, and in 2025 the VENTANA TROP2 (搜索) RxDx device incorporating QCS received FDA Breakthrough Device Designation, the first such designation for a computational pathology companion diagnostic.
He also describes a genomics-driven AI model trained using bi-allelic CDH1 inactivation as an orthogonal ground truth for invasive lobular carcinoma (搜索). "The model classified the disease with high accuracy, but the most informative observations emerged from cases predicted to harbor CDH1 inactivation that lacked the expected coding alterations," he wrote, noting these discordances were enriched for alternative mechanisms including a deleterious fusion, non-coding alterations, and promoter methylation.
The Path to Biomedical Superintelligence
Dr. Reis-Filho defines biomedical superintelligence in an operational sense as "a system that generates novel biomedical knowledge that withstands empirical validation and translates this knowledge into better decisions for patients than the best-qualified and/or world-leading human experts can make when given access to the same information."
He emphasizes that frontier models are necessary yet insufficient on their own. "The system built around the models: the lab-in-the-loop infrastructure, the provenance-rich evidence base, the encoded judgment of a field, the evaluation harness, and the fit-for-purpose model selection and training, will become the durable moats that persist across model generations and appreciate with use," he wrote.
He also cautions against what he terms "AI hypothesis slop," warning that without incorporating negative and null results as model-native data, "autonomous systems will generate hypotheses faster than science can eliminate weak ones." He advocates for a "human on the loop" model, in which a human supervises the run rather than executing it, entering after verification "to make the call that spends irreversible resources... on the strength of a verified record rather than the system's self-report."
"Their ultimate measure will be the validated discoveries that change a patient's trajectory by enabling the right diagnosis, the right biomarker and the optimal treatment at the correct time," Dr. Reis-Filho concluded.