About
Nikolay Nikolov
Artificial Intelligence researcher and engineer
My work is at the intersection of Reinforcement Learning, Computer Vision and LLMs. I am particularly interested in learning multimodal behaviors through these methods and applying them to Robotics and World Modeling.
Currently, I am working on AI projects in the areas of Generative AI and vision-language-action (VLA) models.
Previously, I was one of the first AI Applied Scientists at Wayve (autonomous driving unicorn), where I worked on foundational models and developed the first offline RL method for autonomous driving capable of handling the streets of central London.
Graduated with an MEng in Computer Science and Electronic Engineering from ETH Zurich and Imperial College London, where I was advised by Prof. Andreas Krause and Prof. Stefan Leutenegger.
Publications
Information-Directed Exploration for Deep Reinforcement Learning.
International Conference on Learning Representations (ICLR) 2019.
Selected Projects
Architecture rendering with Diffusion models
Research Project
Breaking causal confusion in autonomous driving
Research at Wayve
Offline Reinforcement Learning for driving in London
Research at Wayve
Learning diverse driving skills via data manipulation
Research at Wayve
Learning to drive with RL from human feedback
Research at Wayve
Driving like a human with Imitation Learning
Research at Wayve
Deep Exploration in RL
Research at ETH Zurich
Distributional Deep RL
Research at ETH Zurich
Deep RL in TensorFlow and gym
Open Source Library
Deep RL for Robot Picking
Internship Project at Ocado Technology
Bayesian Fusion for SLAM
Research at Imperial College
Autonomous robot navigation and manipulation.
Robotics Project
C90 to MIPS Compiler
Compilers Project