Scaling Post-training for LLMs
SPOT: a workshop on the principles and practice of post-training at scale.
03 / Gagan Jain
Service & talks
Organizing
Virtual Research Symposium for Students
NTU Singapore & IIT Bombay
Peer review
2025–present
NeurIPS · ICML · ICLR · AISTATS · CVPR · AAMAS
Teaching & mentorship
IIT BombayTeaching assistant
- Intelligent & Learning Agents
- Machine Learning
- Electricity & Magnetism
Student mentorship
Mentored students in machine learning, robotics, and computer vision.
SeDriCa team lead
Led IIT Bombay’s autonomous driving team.
Selected talks
Efficiency for LLMs
Where does compute go and how to be smart about it
WADLA 4.0 · IIITDM Jabalpur
Breaking into ML Research
Analytics Club · IIT Bombay
Inference-time Efficiency in Large Generative Models
- Dec 2024 UCLA Artificial General Intelligence Lab
- Oct 2024 MEDAL Lab · IIT Bombay
Robust Tracking using Model Predictive Control for Self-Driving Cars
Virtual Research Symposium for Students
NTU Singapore & IIT Bombay
News archive
Sep 2026Joined Proximal to build RL environments and post-train models.
Jun 2026New preprint: Adaptive Block Diffusion, on the training–inference mismatch in diffusion language models.
Apr 2026Co-organizer of SPOT: Scaling Post-training for LLMs at ICLR 2026.
Dec 2025Talk on efficient computation for LLMs at WADLA 4.0, IIITDM Jabalpur.
May 2025Joined Microsoft AI as an Applied Scientist II.