This page collects the work I do outside of my main role: public talks, interviews, Kaggle, publications, mentoring, community work, and selected contributions to the ML ecosystem.
Publications & Engineering Posts
Kaggle
Top Notebooks
| Notebook | Competition | Upvotes | Views |
|---|---|---|---|
| EDA and Models | IEEE-CIS Fraud Detection | 2900+ | 199k+ |
| EDA, FE and Everything | Two Sigma Financial News | 2300+ | 170k+ |
| Segmentation in PyTorch | Cloud Images | 1700+ | 90k+ |
| Exploration Step by Step | PetFinder | 1100+ | 60k+ |
| EDA, FE and Interpretation | TMDB Box Office | 1100+ | 64k+ |
| How to Not Overfit? | Don't Overfit! II | 900+ | 47k+ |
Selected Competitions
| Competition | Result | Medal |
|---|---|---|
| Predicting Molecular Properties | 8 / 2737 | Gold |
| Home Credit Default Risk | 64 / 7176 | Silver |
| Santander Value Prediction | 108 / 4464 | Silver |
| TGS Salt Identification | 100 / 3219 | Silver |
| Microsoft Malware Prediction | 120 / 2410 | Silver |
Featured Talks
All 20+ talks (by year)
2025: Journey from Non-tech to Meta Engineer
2023: Fraud Detection via Face Recognition · Pre-authorization (RU) · Medical Chat-Bot
2022: NL-Augmenter (RU) · Data Visualization · Chat-bot story (RU) · 10 years in IT (RU) · Video Super-Resolution
2021: Business to DL (RU) · Data Viz (RU) · Data Viz (EN) · NER practical aspects (RU) · Reusable DL pipelines · Hydra config (RU) · Podcast: Kaggle & ML · DS Career
2020: Digit recognition & topic modelling · Pair Programming: Drug Classification · My Kaggle Story · MoA Prediction · Metro Visualization · DL Pipelines · PyTorch Lightning + Hydra
2019: Gold solution: Molecular Properties · 2019 in Kaggle (RU) · Chaotic client actions (RU) · Data visualization storytelling (RU)
Interviews & Podcasts
All interviews
Community & Mentoring
- Moscow DataFest 6 (2019) — organized "EDA & Visualization" track
- ODS Awards 2019 — best talk in ML Trainings + competition progress
- Moscow DataFest 2020 — co-organized "ML Trainings" track
- ODS Awards 2020 — best talk in Business section + mentoring award
- Completed Hacktoberfest in 2018 and 2020
- Wrote for a week in dsunderhood
- Core contributor to news-viz — open-source topic modelling on news