Schlagwörter
Skills
- Programming Languages: Python, SQL, PL/SQL
- Deep Learning: PyTorch, Keras, TensorFlow
- Databases: Oracle, Postgres
- Hadoop: Hive, Apache Spark, Impala
- NoSQL: Redis, MongoDB
- Other: AWS, Docker , Airflow, Elasticsearch, bash scripting, Linux, git
- Minor programming skills: R, C, C++, Scala, HTML/CSS/JS
Interests:
- Machine Learning
- Recommendation systems
- Personalization
- Search relevance
- Natural Language processing
Over 10 years' experience in Data Science, Big Data and Business Intelligence. I have worked on innovative projects in e-commerce, retail, telecommunications. I enjoy solving complex analytical problems and developing smart services based on AI and machine learning. I am proud that my work helps to improve the shopping experience of millions of people.
Projekthistorie
Sportmaster is one of the three largest sporting goods retailers in Europe with 210 million offline visitors a year and a huge online customer base.
- Working on improving search relevance and increasing financial result of the search engine using ML approaches
- Developed and implemented ML-based autosuggest API (phrases, categories, items)
- Developed and implemented automated pipeline of dataset labeling for assessing search relevance using crowdsourcing platform
- Developed and implemented item ranking personalization in catalog branches on site and mobile app
- Developed and implemented system for optimizing marketing communications using uplift models
- Developed and implemented recommender systems based on Deep Learning approaches
- Implemented pipeline for preparing product embeddings based on purchases, text descriptions, images
- Implemented pipeline for extracting product tags using CV, NLP (NER)
- Conducted A/B tests before implementations
MegaFon is the second largest mobile operator in Russia with 75.4 million customers.
Worked on solving business problems and implementing prototypes:
- Improved the Churn prediction model
- Developed predictive model for client gender
- Developed classification model for Down-sale prediction
- Prepared various triggers to stimulate top up, prevent the churn
Aeroflot (the largest airline in Russia).
- Unified client profile model (UCP) - comparison of user data from different sources (Oracle DWH, GA, Siebel, Check-In system, social networks) using near-duplicate detection algorithms
- Client LTV forecasting
- RFM segmentation
- Developed and implemented recommender system for email personalization
SAP BI implementation projects in Federal Grid Company (operator and manager of unified electricity transmission grid system in Russia), MOESK (major distribution electric grid company in Russia)
- Implemented wide range of management reports for desktop and mobile devices to support decision-making in investment activities
- Reporting, Data warehouse modeling (SAP BW), ETL, ABAP development
SAP BI implementation projects in FINN FLARE, Incity, SELA (large Fashion Retail companies in Russia), Euroset (large mobile phone retailer)
- Implemented management reports and KPI dashboards for managers of different levels, Data warehouse modeling (SAP BW), ETL