Kirill Danilyuk
I do perception for robots. I assembled and led a team of computer vision engineers at Yandex Self-Driving Cars (now Avride); we shipped from anomaly detection on sensors to production monocular depth for robots. I am also a private pilot so I am seeking projects in the intersection of robotics and aviation. I am currently exploring 3D vision and VLM / VLAs for robotics and autonomous drones.
Projects
- SLAM Wand - I'm currently working on an AI stick that maps my house in 3D, localizes me, and makes a nice 3D render.
Publications
- March 2025 (article for The Insider) DeepSeek vs. ChatGPT: China joins the AI race against the U.S., but America has a response
- April 2018 (post for TDS) ConvNets Series. Actual Project Prototyping with Mask R-CNN
- Sep 2017 (post for TDS) ConvNets Series. Spatial Transformer Networks
- July 2017 (post for TDS) ConvNets Series. Image Processing: Tools of the Trade
- Feb 2017 (post for TDS) CarND Project 1: Lane Lines Detection — A Complete Pipeline
Activities
Nebius Academy — Generative AI Course Team
I was a course team member focusing on the diffusion models. I presented a lecture, a practical workshop on implementing diffusion models, and a homework assignment.2024
Yandex Data School — Self-Driving Cars Course
Main contributor to the Perception class for the Self-Driving Cars course, covering computer vision and sensor fusion topics.2022
Habr Interview: Data Science Career Insights
Interview discussing the journey into data science, career advice, and insights into the field of data science and machine learning.2018
Jigsaw: Toxic Comment Classification Challenge — Top 1% Score (Silver Medal)
We developed a mini framework to facilitate and automate building data pipelines, prepared several dataset preprocessings and combined 40+ L1-models in a single solution.2018
Led Data Science Training for BNP Paribas (Luxembourg)
Together with New Professions Lab, we conducted a two-month data science course for BNP Paribas employees in Luxembourg.2017
Allstate Claims Severity — Silver Medal Solution
My solution to Allstate Claims Severity challenge that won a silver medal on Kaggle, featuring advanced regression techniques and feature engineering.2016
Talks
Podcast: How Self-Driving Cars Work
We discussed my experiences and insights in the field of autonomous vehicles at Yandex SDC.2022
Sirius.Science: A lecture on Self-Driving Cars for gifted children in Sochi
I shared insights about creating autonomous vehicles, discussing key challenges like precise navigation, sensor technology, and road environment reconstruction.2019
PyData Moscow — Semi-Automatic Data Labeling Pipeline
End-to-end flow for building a scalable semi-automatic labeling process in production. Practical patterns, pitfalls, and ROI.2018
Road Sign Classification — STN + IDSIA, ipyparallel, TensorFlow
A practical walkthrough of robust traffic-sign recognition with spatial transformer networks.2017