An engineer with a PhD in AI/Robotics with an expertise on visual control of robotic systems with deep reinforcement learning.
My name is Sardor Israilov and I was born on the 05/04/1997 in Uzbekistan. I am an engineer, living in France for more than 8 years. I have a PhD in AI/Robotics with the focus on visual control with deep reinforcement learning. During my PhD I had hands-on experience with deep learning methods applied on 3 different robotic systems: cart-pole, static robotic fish and moving robotic fish. My work incorporated deepened knowledge of deep RL applied on the control of robotic fish and appropriate physical derivations. You can find more information about my PhD in my thesis.
Prior to Phd, I worked as an engineer-apprentice for 3 years at Solystic and I had my Master's degree from INSA Lyon. Working at Solystic allowed me to acquire solid skills in programming at Matlab/Simulink. As my final year project at INSA Lyon, I implemented Mask R-CNN to segment automatically medical images; the model was trained on the cloud.
At some point, I was engaged in robotics club, where I had a chance to participate in France Robotics competition (CdF robotique). I am a lifelong learner and I completed numerous online courses including Udacity Robotics and ML nanodegrees where I had a chance to program in python and C++ for projects like “home robot(SLAM) in gazebo”, “Deep RL robotic arm”, “point cloud” etc. available in my github. Finally, I am passionate about cutting-edge AI techniques having big impact on our daily lives and I want to be part of it.
Check out preliminary paper: deepFish
Some hardware developed overview: overview
Swimming gait optimization of a robotic fish via Deep Reinforcement learning using camera (CNN), force sensors (FCN); work with sensors such as IMU, LIDAR...
Check out the video: article: Video of deep RL
Another video of model-based control: LQR and Lyapunov
