Path of Data & AI Challenges
summary
Radiology workflows move from paper archives to one‑click digital access, enabling AI to cut reporting backlog but raising patient privacy and data curation work; a small team builds Lawabot: Serbian‑law Q&A with article citation, fine‑tuned because base LLMs don’t know local statutes and deployed successfully on a demo booth. Large‑scale work focuses on pipelines, post‑training and inference cost — hosting 400B‑parameter models across 8–16 GPUs and optimizing kernels with Triton/PyTorch to turn minute‑long video generation into seconds.
AI-written summary · about
chapters (11)
- 0:00Introduction and welcome
- 1:15Radiologist’s data journey
- 12:21AI challenges in healthcare
- 18:52Legal AI project overview
- 23:51Collaboration in AI projects
- 27:31Open source in AI ecosystem
- 30:54Legal tech and AI regulations
- 34:06Education and knowledge sharing
- 37:20High performance computing needs
- 41:00Cross-domain collaboration advice
- 44:37Q&A and networking
related talks
09mHow AI is changing the healthcare system
28mThe Future of Healthcare – AI’s Promise and the Challenges Ahead
36mResponsible AI
13mBuilding High-Performance AI Literature Processing
50mA Sustainable Future for Learning AI in Healthcare
06mTransitioning AI from the lab to real-life applications
17mAI/ML demo derby
23mAI in the Life Science Sector