A Data Science Startup: Let It Wave's Meandering Adventure and Its Duller Sisters

Explore the journey of a data science startup in this 32-minute talk by Jérôme Kalifa at the Institut des Hautes Etudes Scientifiques (IHES). Trace the evolution of Let It Wave from its founding in 2001 by Stéphane Mallat and three students to its acquisition in 2008. Discover how the company transformed from initial consulting activities to an ambitious fabless venture in consumer electronics for HDTV. Gain insights into the challenges and successes of a startup originating from academic research, specifically Le Pennec's thesis on bandlets. Learn about the company's ventures into seismic inversion and stratigraphy for oil exploration, as well as its transition to a VC-funded fabless operation focusing on frame rate conversion and chip design. Compare Let It Wave's fast-paced journey with Kalifa's subsequent ventures, including Lixoft, which focuses on modeling and simulation of clinical trials. Examine the pros and cons of non-startup approaches and explore data science challenges in observing mental illness through the Let it Care project. This talk offers valuable lessons for aspiring entrepreneurs and researchers interested in the intersection of academic research and business innovation.

Intro
2001 - Le Pennec's thesis on bandlets
2001-2004 Services for industry applications
Seismic inversion and stratigraphy for oil exploration
2006 HDTV and the transition to a VC-funded fabless
Step 2 - Frame Rate Conversion
Chip Design and cost-competitiveness
Lixoft, modelling and simulation of clinical trials
Pros and cons of a non-startup
Example on observing a mental illness
Let it Care data science challenges