• PyTorch • Reinforcement Learning • NumPy • Pandas • MLflow • Weights & Biases • ONNX • ROS2
• Python • C++ • Docker • Git
• Graph Algorithms • Dynamic Programming • Markov Decision Processes (MDP) • Path Planning

• Developed intelligent resource orchestration for VBS/LTE networks using Reinforcement Learning within the ns-3 simulation framework.
• Explore the research on Zenodo

• Authors: Louis Nisiotis, Nikita Markov
• Architected a deployment pipeline that achieved a 2.3× model size reduction (from 1.1B parameters down to 470M) for Small Language Models (SLMs) integrated into virtual worlds.
• Leveraged TVM, MLC-LLM, and multi-scheme quantization (INT3–INT8) to optimize performance via ONNX and TensorRT (trtexec) directly on hardware-constrained NVIDIA Jetson edge devices. Explore

I am captivated by systems that close the loop between perception and action. My primary research and development interests lie in:

• Autonomous Driving: Mapping out safe, optimal trajectories through complex, dynamic environments.
• Robotics: Giving physical agents the computational brainpower to reason about their surroundings.
• Edge AI: Squeezing maximum intelligence out of limited hardware, because AI belongs where the action happens.