Nauman Ahmed

Multi-Agent 3D Coverage Path Planning via Hybrid LTC-PPO for UAVs • Developed LTC-based MARL for 3D coverage under terrain, NFZ, and energy constraints. • Created voxel simulation (20×20×6) with collision avoidance and BFS-based forced-return. • Integrated PPO with LTC and attention for stable policy learning and action masking. • Trained for ~11K episodes; achieved ~93% coverage across 100 unseen environments. • Built diagnostic pipeline to optimize KL divergence, entropy, and safety.

• Built 5+ high-performance React.js apps with a mobile-first focus. • Boosted IoT platform speed by 35% via code-splitting, state management, and rendering optimization. • Managed full SDLC from UI/UX wireframing to deployment, ensuring accessibility and cross-browser standards. • Reduced technical debt through rigorous debugging, testing, and workflow documentation. • Developed reusable component libraries and modern frontend tooling to significantly accelerate development timelines.