Jelin Raphael Akkara
I am pursuing a PhD in the BMCS programme at the University of Padua (UniPD), in collaboration with the Fondazione Bruno Kessler (FBK). My research in Embodied AI investigates how autonomous agents can understand complex environments, identify the causes of failure, and adapt their behaviour to recover robustly. I am particularly interested in multimodal reasoning, vision-language models, memory, personalization, and causal and counterfactual approaches to intelligent interaction in 3D environments.
News
- [September 2026] — Presenting IMPRINT at IROS 2026
- [September 2026] — Organizing Committee Member for ICANN 2026 at Padua, Italy
- [July 2026] — Presented PersONAL at ICRA 2026
- [July 2026] — PersONAL-3D published at IJCV!
Teaching Assistant – Computer Vision and Cognitive Systems
University of Padua
Oct-Dec 2025
Assisted in the course "Computer Vision and Cognitive Systems". Responsibilities included conducting lab sessions, providing guidance to students on assignments and projects, and facilitating discussions on advanced topics in computer vision and cognitive systems.
GenAI Research Intern
ViWear
Jul-Oct 2025
Enhanced a Stable Diffusion-derived virtual try-on model, improving perceptual metrics (FID, KID, LPIPS) by 7-12%, and tripling the inference speed from 1 fps to 3 fps. Implemented architectural improvements, paired with GPU-constrained fine-tuning, resulting in more realistic try-on results while preserving both target cloth and pose authenticity.
Computer Vision Research Intern
Fondazione Bruno Kessler (FBK) /
Visual Intelligence and Machine Perception Group (VIMP), University of Padua
Oct-Mar 2024
Guided by: Prof. Lamberto Ballan, Tommaso Campari
My work focused on enhancing scene mapping using multi-modal (Vision-Language) features, while keeping the pipeline memory-efficient. This resulted in a 7.6% increase in navigation success rate while achieving a 99.75% reduction in memory usage. I was also involved in developing a modular framework for easily accessible mapping techniques in Object Goal Navigation.
PhD in Brain, Mind, and Computer Science (BMCS)
University of Padua (UniPD) / Fondazione Bruno Kessler (FBK)
2025-(Present)
Research Focus: Embodied AI, Multimodal Reasoning, Vision-Language Models, Memory, Personalization, Causal and Counterfactual Approaches to Intelligent Interaction in 3D Environments
MSc in Physics of Data
University of Padua, Italy
2022-2025
GPA: 110 / 110
Thesis Title: Zero-Shot Object Goal Navigation using Online Image Retrieval
B.Tech in Engineering Physics
National Institute of Technology, Calicut, India
2018-2022
GPA : 8.43 / 10
First Class with Distinction