SBER Risk Modeling & Research:
- DS Intern (May 2023 - October 2023)
- Junior DS (November 2023 - December 2023)
- Junior+ DS (January 2024 - December 2024)
- Middle DS (January 2025 - present day)
Tasks:
- Built 3 versions of Risk driven & Profit-Loss (RPL) models for decision-making on consumer loans
- Built several transactional neural networks (for transactions of length 800 and 5000):
- SOTA multitask model using transofrmer-based approaches
- GRU x GigaChat model
- Linformer model (self-attention with linear complexity)
- Performed calibration and modeling of the migration matrix for various versions of RPL models
- Wrote a library for automating calibrations of RPL models - the time was reduced from 2 days to 15 minutes
- Carried out feature engineering of the credit portfolio
- Performed ad-hoc analytics on developing models
- Conducted several interviews for internship positions
- NUST MISIS Applied Mathematics (2021 - 2025)
- Deep Learning School (2021 - 2022) | Diploma
- 1/400 🥇 Yandex AIDAO (AI International Olympiad | Online stage) (2024) | Solution
- 1/20 🥇 IT Purple Hack "Churn of the SBER's salary client" (2024) | Diploma | Solution
- 3/20 🥉 International Digital Breakthrough "Automatic dispatch of applications" (2024) | Diploma | Solution
- 3/20 🥉 International Digital Breakthrough "Churn of customers in the fast food industry" (2023) | Diploma | Solution
- 3/20 🥉 Digital Breakthrough "Prediction of the purchase of a parking space" (2023) | Diploma | Solution
- ML & DL Models From Scratch
- Face Recognition
- RNN Name Generator
- Semantic Segmentation of Cancer Cells
- ⚽ Football
- 🏋️ Gym
- ⛷️ Skiing
