Bonjour · Servus · Vanakkam
I’m Lalith — Postdoctoral Researcher at CAMMA, IHU Strasbourg & University of Strasbourg.
My mission is to build the next generation of safe and interpretable AI systems for surgery by combining modern computer vision with real‑world surgical data.
What I do
I work with large‑scale, real‑world, heterogenous surgical data to build models designed to:
- Operate across centres, procedures, and devices.
- Use vision–language pretraining & foundation models tailored to surgical video semantics.
- Detect rare but critical intra‑operative events.
- Enable safe and human-in-the loop decision support.
Current focus
- Rare event detection
- Action anticipation
- Temporal event localization
- Surgical foundation models
- Vision–language pretraining
Track record
- 15+ publications in high-impact journals and conferences (IEEE JBHI, MICCAI, EJCTS, Annals of Thoracis Surgery)
- Reviewer CVPR · NeurIPS · MICCAI · IEEE TMI · MedIA
- 4× poster awards
- 8+ projects mentored
Beyond the lab
Scientists shouldn’t be hermits. I’m a passionate science communicator and educator with 100+ hours of teaching. Through science slams across Germany, I’ve shared my work with an audience of 5,000+, earning 3× best slam awards and 2× semi‑final nominations at the German Science Slam Championships.
News
-
I joined the CAMMA Group, IHU Strasbourg as a Postdoctoral researcher, where I will be working on CompSURG, an ERC-funded project for large-scale multi-centric multi-pocedural surgical video analysis.
-
I successfully defended my doctoral degree (Dr. sc. hum.) and graduated from Heidelberg University. My cumulative dissertation titled Deep learning based image analysis for endoscopic minimally invasive mitral valve repair was awarded the highest grade of summa cum laude.
-
I delivered two science slams: in Heidelberg (🥇 first place) and Mainz (🥇 shared first place).
