Research Interests
- Graph Representation Learning (GNN Architectures, Expressivity)
- Over-Squashing And Over-Smoothing In Graph Neural Networks
- Continual Deep Learning
- Continual Reinforcement Learning
- World Models And Model-Based Approaches
- Model-Free Reinforcement Learning And Sample Efficiency
Recent Work
-
Nov 2025 — Value & Policy Learning (SAC, JAX)
Implemented Soft Actor-Critic with JAX/Brax/Equinox, including uv-managed deps, training loop, and interactive Plotly returns. See project: value_n_policy_learning. -
Nov 2025 — Microwave Imaging with CNO
Complex-valued scattering simulation and Convolutional Neural Operator-based reconstruction; results and metrics for ID/OOD generalization. Read the write-up: Microwave Imaging (CNO).
Research work — coming soon
I'm preparing research code, notes, and experiments. Check the research page for updates and drafts.