Talks & podcasts
I regularly speak at podcasts, meetups and conferences (mostly in Russian) about practical DS/ML and AI — from MedTech and anti‑spoofing to football analytics and AI project failures. Below are some of the talks.
Podcast “Code & Coffee #8 – Deep Learning & COVID‑19”
A conversation about my experience as a Deep Learning Scientist in a Kazakh MedTech project:
– how deep learning models help diagnose COVID‑19 pneumonia;
– what kind of data and processes are needed;
– what MedTech monetization looks like in reality.
nFactorial Talks – Data Science, Big Data & careers
We discuss DS/ML career paths, big data, moving from individual contributor to Head of DS, and why thinking in terms of products and business is crucial for data scientists.
Kolesa Conf 2021 – Face Anti‑Spoofing
A talk on face anti‑spoofing for biometric systems:
– attack types (photos, videos, masks, 3D models, deepfakes);
– limitations of classical methods;
– neural approaches and architecture of a production anti‑spoofing service.
DSMLKZ – All about football data analysis
A talk on football analytics and working with data in sports:
– fitness/GPS vs event data;
– visualising games and player actions;
– metrics like xG, PPDA, EPV;
– modeling team styles and using NMF to understand and compare teams.
AI Club – nobody talks about AI failures
An honest breakdown of typical AI/ML project failures:
– non‑working MVPs that live only in slides;
– over‑rated metrics and models without real‑world value;
– what needs to change so that AI projects stop being “slide‑ware” and start delivering.
beetech conf by QazCode
Participation and talks at beetech conf by QazCode as Big Data curator and speaker in AI & Beyond streams — about practical AI adoption in telecom, data‑driven products and building a healthy DS/ML community.
Conference website: https://beetech.kz
Most of these talks are in Russian. For a structured overview of my public speaking and links to related blog posts, you can also check the Russian page “Подкасты и доклады”.