My recent work explores generative 3D world models, specifically diffusion-based occupancy prediction for exploration, mapping, and long-horizon planning in mobile robots. I develop algorithms and AI models that enable mobile robots to perceive, predict, and act effectively in complex, partially observed, and egocentric environments. My work focuses on: multi-modal machine perception for 3D scene understanding, generative occupancy modeling, and Vision–Language–Action (VLA) models that unify perception, language, and control for grounded decision-making.

---

Publications

RF-Modulated Adaptive Communication Improves Multi-Agent Robotic Exploration

• In Review - 2026
• Read our pre-print manuscript here.

Robust Robotic Exploration and Mapping Using Generative Occupancy Map Synthesis

• Accepted at Autonomous Robots (Springer Journal) - January 2026
• Springer Autonomous Robots Journal.
• Check out the pre-print manuscript here.

Online Diffusion-Based 3D Occupancy Prediction at the Frontier with Probabilistic Map Reconciliation

• Accepted to ICRA 2025: Check out the pre-print manuscript here.
• Read the published version on IEEE Xplore
• Paper Webpage

SceneSense: Diffusion Models for 3D Occupancy Synthesis from Partial Observation

• Accepted to IROS 2024: Read our pre-print manuscript here.
• Read the published version on IEEE Xplore
• Paper Webpage

---

Projects and Presentations

Advancing Robotics with Vision-Language-Action Models

• Watch my talk here

Sim-to-Sim Transfer Framework

• Developed a simulation-to-simulation transfer pipeline to validate control policies across different physics engines (comparing IsaacSim with MuJoCo).
• Read the project write-up here

Vision Language Models: PaliGemma from Scratch

• Recreated the PaliGemma Vision-Language Model (VLM) architecture entirely from scratch using PyTorch.
• Implemented the complete model structure to deepen understanding of multimodal integration and large-scale model design.
• View the GitHub repo here