Think Forward.

Tariq Daouda Podcast

by Tariq Daouda
2539
Chapters: 2 1.0 min read
This series recapitulates all episodes of my podcast. It focuses on state-of-the-art research or deeptech, with researchers, startup founders and stakeholders of the innovation ecosystem. Every episode goes deep into a topic or scientific paper.

1: Exposomics: Lifelong Environmental Exposures and Health Outcomes with Dr. Jacob Galan 2463

Tariq Daouda, 20 d.
During this conversation we explore the concept of the exposome, which encompasses all environmental exposures a person experiences over a lifetime. The discussion covers how mass spectrometry helps analyze these exposures and sheds light on gene-environment interactions contributing to health outcomes. We examine how environmental toxins are linked to chronic diseases and highlight how cultural and dietary habits shape individual exposure profiles. The conversation also discusses how artificial intelligence is transforming exposomic research, offering the potential for new health insights through advanced analytics. We address the complexities involved in studying exposomics: the significance of protein structures, innovations in biological research methods, and improvements in mass spectrometry that provide richer data for analysis. Challenges such as finding exposomic markers and detecting elusive ‘ghost molecules’ are considered, as well as the interactions between environment and hormonal health—like changes in testosterone levels. They discuss the promising future of exposomics, focusing on integrating AI and biomonitoring to better understand and manage the real-life impact of environmental exposures on health.
youtube.com/watch?v=6L7j8UdTgGA

2: Agentic AI Beyond Benchmarks: Meta-Agents & the Future of AI Evaluation with Khalil Mrini 76

Tariq Daouda, 3 h.
I recently sat down with Khalil Mrini to talk about his work and international experiences. He has spent time in Marrakech, Switzerland, India, and the United States, each place influencing his perspective in different ways. We also mentioned his visit at the UM6P, his experience of the university, students and innkvative AI curriculum. Khalil presented his new paper on agentic AI. The paper focuses on the use of autonomous agents to evaluate and benchmark other agents: essentially, systems that can test one another’s capabilities. He described how this approach could provide a more dynamic and optimal method for measuring progress in AI research. We ended the conversation by discussing AI ethics. Our exchange raised open questions about responsibility, transparency, and how the field can ensure that increasingly autonomous systems align with human values.
youtu.be/zE7PKRjrid4