Krishnam Soni
I'm interested in machine learning and world models. From May to September 2026, I was a research intern at the CPS Lab, Montanuniversität Leoben, working on world models for robotics. My main focus was test-time adaptation: correcting a pretrained world model when a robot's dynamics change, without retraining it. The internship led to three papers: one submitted to ICRA 2027 and two accepted at NeurIPS 2026 workshops.
Before that, I worked on medical image segmentation: 3D cardiac MRI at NTU Singapore, and retinal OCT scans at IIT Kharagpur, which led to a paper at IEEE INDICON 2024. I completed my B.Tech in Computer Science (AI & ML) at Manipal University Jaipur in 2026, which included an exchange semester at Montanuniversität Leoben.
Research interests: Machine learning World models Test-time adaptation Computer vision
News
- Oct 2026ALRA and PULSE accepted at NeurIPS 2026 workshops.
- Sep 2026Sandwich-Residuals is on arXiv and submitted to ICRA 2027.
- Sep 2026Reviewing for the NeurIPS 2026 WM PAI and TTCL workshops.
- May 2026Started a research internship at the CPS Lab, Montanuniversität Leoben, Austria.
- Oct 2025Began an exchange semester in Industrial Data Science at Montanuniversität Leoben on a MULgraIN scholarship.
- May 2025Started a remote research internship at NTU Singapore on cardiac MRI segmentation.
- Dec 2024OCT segmentation paper accepted at IEEE INDICON 2024.
- Jun 2024Started a research internship at IIT Kharagpur on OCT scan segmentation.
Selected research
All publications →
