Leandro Agudelo holds an MSc in Biochemistry from Stockholm University and a Medicine Doktor and Ph.D. degree from Karolinska Institute, Sweden, where he specialized in machine learning engineering and bioengineering applied to dynamic complex systems, studying resilient strategies in biological systems for genomics and metabolism. Following a postdoctoral research at the MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Broad Institute of MIT & Harvard, he became a Research Scientist and Principal investigator at MIT-CSAIL with supporting funds from Novo Nordisk. By the end of 2024, Leandro transitioned from academic tenure-track to found Daice Labs, pursuing his scientific vision while translating breakthrough research into practical applications.
His research focuses on computing natural intelligence principles for resilient and adaptive systems. His work combines two complementary areas: developing composite AI architectures (cellular-inspired systems with adaptive optimization and resilient strategies) and their application across complex domains (bioengineering, dynamic systems). This approach bridges fundamental research in computing biological intelligence with practical solutions for evolving systems.
His research excellence is recognized through multiple honors including Karolinska institute and MIT Ph.D. and Postdoctoral Fellowships, Swedish Research grants, and more than 2 million in funding from industry partnerships including Novo Nordisk. He was nominated for best PhD thesis at Karolinska institute, and has received multiple scientific presentations awards. His scientific contributions include publications in leading journals such as Cell, Cell metabolism, Science, Nature communications among others. Beyond academia, he has successfully launched ventures in real estate and private equity investments. By the end of 2024, he assembled an interdisciplinary team at Daice labs, securing preseed funding and strategic partnerships.
At Daice Labs, Leandro builds upon his scientific foundations from MIT, developing frameworks to compute natural intelligence into composite AI architectures. His unique interdisciplinary background in biological systems, bioengineering, and AI guides our vision of decoding and implementing resilient strategies inspired by cellular sytems (e.g., multicellular collaboration, cooperative specialization, adaptive organization, division of complementary function). We believe that the future of computing requires evolutionary-like infrastructure to optimize growing complexity in multi-modal and multi-agentic systems. Under his leadership, Daice Labs pioneers systems-inspired approaches to understand and modulate complex domains, from finance and digital commerce, to techbio and AI innovation.
Summary: A multidomain computational and experimental framework reveals how cellular specialization emerges through adaptive compartmentalization during low-energy states, establishing parallel principles for both biological systems and adaptive AI composite systems (§ supervised this work).
Summary: Metabolic HAR-genomic hubs regulate energy conservation and adaptation through dynamic transcriptional organization, revealing biological computational principles that can enhance the development of resilient and adaptive systems
(§ supervised this work).
Summary: NK cells demonstrate how molecular repair hubs coordinate stress response and adaptation during energy conservation, revealing natural principles for engineering and computing systems with built-in repair mechanisms and adaptive memory under resource constraints (§ supervised this work).
Summary: We demonstrate that non-equilibrium tuning and compartmentalization, modeled using operads, are sufficient to replicate cellular intelligence processes—such as specialization and communication—and implement this framework as IntCyt, an unsupervised learning algorithm that surpasses conventional models in data mining through generative and self-supervised tasks (§ supervised this work).
Summary: We used a multi-omic integration approach to discover resilient strategies induced by exercise in skeletal muscle. PGC-1α1 activates a metabolic mechanism in exercised muscle that connects kynurenine metabolism to the malate-aspartate shuttle, enhancing bioenergetic efficiency and exercise performance.
Summary: The tryptophan-kynurenine pathway mediates interorgan communication through metabolites that affect immune, inflammatory, and neural processes, offering potential therapeutic targets for resilience as they are modifiable through exercise and lifestyle interventions.
Summary: Exercise-induced muscle kynurenine metabolism produces kynurenic acid that activates Gpr35 in adipose tissue, enhancing energy utilization and metabolic health through increased lipid metabolism and thermogenesis, revealing a therapeutic pathway for metabolic disorders.
Summary: Computational analysis of molecular signatures induced by exercise in skeletal muscle reveal adaptations driving resilience to stress. Exercise protects against depression by activating muscle PGC-1α1, which increases kynurenine aminotransferases to convert depression-inducing kynurenine into kynurenic acid that cannot enter the brain, revealing a peripheral therapeutic target for stress-induced depression.
"Where nature's resilence meets intelligent systems"
© 2024 Daice Labs Inc.
