PhD Researcher, Indian Institute of Science, Bengaluru

AI for Medical Imaging · Representation Learning · Self-Supervised Learning · Reliable Healthcare AI

I work on machine learning methods for healthcare and medical imaging, especially in settings where labels are limited, expensive, noisy, or unavailable. My research interests include self-supervised learning, structured latent representations, calibrated risk estimation, and deployment-aware evaluation of AI systems for clinical workflows.

• Self-supervised and representation learning for medical imaging
• Data-efficient learning under limited or unlabeled data
• Autoencoders and structured latent spaces
• Calibration, risk estimation, and healthcare workflow simulation
• Reproducible ML pipelines for clinical AI research

• Two-Fold Patch Perturbation for Efficient Self-Supervised Learning in 3D Medical Imaging IJCAI-ECAI 2026

• Calibrated Risk and Queueing Simulation for AI-Assisted Radiology Worklist Triage DAI & AIMedHealth @ AAAI 2026

• DIME: Deterministic Information Maximizing Autoencoder DeLTa @ ICLR 2025

• Learning Low-Rank Latent Spaces with Simple Deterministic Autoencoder WACV 2024

• DeepVAT: A Self-Supervised Technique for Cluster Assessment for Image Datasets ViPrior @ ICCV 2023

I am gradually organizing research code, project pages, and reproducible pipelines related to medical imaging, representation learning, and healthcare AI. Keep an eye on my page (pinned projects) for exciting and cool updates!!

• Website: https://tirthajit.github.io
• Google Scholar: https://scholar.google.com/citations?user=gDr0VrYAAAAJ
• OpenReview: https://openreview.net/profile?id=~Tirthajit_Baruah1
• ORCID: https://orcid.org/0009-0009-2020-0521
• LinkedIn: https://www.linkedin.com/in/tirthajit-baruah/