Alexander Golubev Email & Phone Number

Amsterdam, NH, NL

Amsterdam, North Holland, Netherlands

San Francisco Bay Area

- Developing a real-time system for products matching from scratch 1. Microservices communication (gRPC, Kafka) 2. Fast and scalable machine learning models (ONNX, Triton Inference Server) 3. Databases for efficient data storage (PostgreSQL, Redis, HBase)- Additionally managing a team of 5 ML engineers

Москва, Россия

- Developed a search suggest ML algorithm which significantly beat the current production solution - Integrated Spark Structured Streaming pipeline in ETL processes- Designed and implemented Prod2vec architecture (Blog post: https://towardsdatascience.com/vector-representation-of-products-prod2vec-how-to-get-rid-of-a-lot-of-embeddings-26265361457c)

Москва, Россия

- Came up with a more relevant metric for model evaluation which speeded up hypothesis testing- Significantly optimized PySpark jobs that eventually saved 20% of working pipeline time- Added detection system into matching pipeline to get more precise embeddings- Developed several DL models and integrated them in PySpark- Did research in metric learning.

Москва, Россия

Paper: https://arxiv.org/abs/2206.06827Blog post: https://towardsdatascience.com/evolution-of-policy-gradient-methods-in-reinforcement-learning-from-reinforce-to-evm-4e80017a297b- Researched RL algorithms, reviewing variance reduction techniques- Implemented new algorithms based on policy gradient methods

Москва, Россия

- Leaded medicine OCR project from problem statement to production- Applied SOTA deep learning models for text detection and recognition- Created several Pytorch models as microservices- Communicated with medical consultants to improve the quality of the model

Москва, Россия

- Developed the churn prediction system allowing to react quickly to sudden layoffs- Aggregated data from databases and structured information flow in processes

Лондон, Англия, Соединённое Королевство

- Developed a mobile application for employees for acquaintance with GDPR- Took courses on accounting and financial tools