"Counting my first 10000 hours in solving AI."
I am a Junior pursuing Computer Science and Engineering at the Indian Institute of Technology, Jodhpur. I love AI research and Math. My main areas of interest are Generative Modeling for Computer Vision, Scene Understanding, and XAI to ensure privacy-respecting multimodal systems. I am extremely grateful to be jointly mentored by Dr. Shivang Agarwal and Dr. Mayank Vatsa.
I am keenly pursuing opportunities to engage in pioneering research alongside Visionary teams committed to AI research-driven innovation.
- July, 2025: Paper accepted to ICCV, U&ME Workshop 🎉!
- Summer, 2025: Research Intern @ TAILS, Thoughtworks.
- Summer, 2025: Research Project Assistant @ UCR. PI - Dr. Yinglun Zhu
- January, 2025: TAing CSL1020 - Introduction to CS, S25. Instructor - Dr. Mayank Vatsa
-
November, 2024: Joining the
Image Analytics and Biometrics Lab
as an undergraduate RA, where I am working on "Machine
Unlearning for Multimodal systems".
PI - Dr. Mayank Vatsa, Co-PI - Dr. Shivang Agarwal - October, 2024: Building flow-based, diffeomorphic generative models for "Fast Shower Simulation (FastSim)" in High Energy Physics for CERN.
-
Summer, 2024: Research Intern @
Center for Digital Agriculture, UIUC. Worked on "Towards Unsupervised Latent
Representations for Cattle Image Segmentation".
Mentor - Dr. Tiago Bresolin. - January, 2024: Assistant Head (AI and ML Events) - Prometeo '24 (IIT Jodhpur's annual, national level, technical festival). Gathered 2000+ registrations, gave away prizes worth INR 100000.
- Summer, 2023: Summer Intern (Machine Learning and Data Analyst) @ Fluxgen Technologies
- February, 2023: Volunteered at Agile India 2023, Bangalore
|
Sai Siddhartha Chary Aylapuram, Veeraraju Elluru, Shivang Agarwal ICCV U&ME Workshop 2025 paper / code / program page Deep neural networks often rely on spurious correlations in training data, resulting in biased or unfair predictions, particularly in safety-critical applications. While conventional bias mitigation methods typically require retraining from scratch or redesigning the data pipeline, recent advances in machine unlearning offer a promising alternative for post-hoc model correction. In this work, we explore the trifecta of efficiency, fairness, and model utility post unlearning via Bias-Aware Machine Unlearning, a paradigm that selectively forgets biased samples or feature representations to address various forms of bias in vision models |
|