Dmitrii Khizbullin's professional page
I am a machine learning and computer vision researcher/engineer specializing in deep learning for visual perception in the fields of autonomous driving, robotics, cloud, and edge computing. I've been leading teams of up to 4 engineers, delivering highly reliable software to production, as well as providing fine-grained internal reports as a result of scientific research, pushing the state-of-the-art in computer vision and meta-learning.
I give lectures to Moscow PhysTech students (#1 tech uni in Russia), lecture about deep learning at OTUS online education, and read a course on autonomous driving at Luxoft Training. I've been mentoring and judging at circa 10 hackathons globally.
My courses
- I am the author of the course "LiDAR Perception for ADAS and Autonomous Driving" that I give at Luxoft and at Moscow Institute of Physics and Technology (github, Luxoft Training, my page at Luxoft)
- I am the author of the course "Tensor compilers for neural network training and inference" that I give at at Moscow Institute of Physics and Technology
My patents
- I am the primary inventor of a world-registered patent "Processing a data stream of scans containing spatial information provided by a 2d or 3d sensor configured to measure distance by using a convolutional neural network"
My articles on Medium and Hackernoon
- How Starship robots see the world
- C++17 structured bindings for more safe, functional code
- How to build a multi-threaded pipeline in C++ with std::async
- RANSAC, OLS, PCA: 3 Ways to Draw a Straight Line Across a Set of Points
- Variational Autoencoders (VAE): How AI Learns Whether Your Eyes Are Open Or Closed
- An Algorithmic Implementation of an Autonomous Driving LiDAR Perception Stack with PCL
My talks
- MSSTAGE 2021: How to launch neural network training on Microsoft Azure
- Webinar at HSE: MuZero: solving Atari with computer vision and reinforcement learning
My pet projects on GitHub
-
Implementation of SSD bounding box detector for Kitti
-
Interactive dataset visualization with Dash @ AWS
-
Experimental OpenCL simulation of 256k-particle elastic relaxation
-
Neural-network-based predictor for area of a rectangle
-
Verilog/FPGA driver of TOF camera for Blackfin DSP
-
Semantic segmentation for aeral imaging
- Empathetic chat bot
Materials in Russian
- [Youtube stream] My talk at Google DevFest Karaganda 2022 about MuZero
- [Youtube stream] My talk at Neconf Karaganda about metrics in machine learning
- [Youtube webinar] Accuracy metrics in machine learning
- [Youtube webinar] Transfer learning by means of interleaved training
- [Youtube webinar] How to start your career in Data Science
- [Youtube podcast] Vision of autonomous cars
- @bespilot
- My page at OTUS online education