Senior Machine Learning Engineer (Technical Co-Lead)
Perseptex
Dec 2025 — PresentCo-leads ML engineering for a contract platform, delivering contract Q&A and voice-agent features with measurable accuracy, workload, and latency gains.
• Achieved 90% answer accuracy in contract question answering through a RAG service combining document ingestion, metadata extraction, Milvus search, and FastAPI. • Built a Gemini Live voice agent with a fine-tuned intent classifier using PyTorch, reducing manual outreach and follow-up workload by 30% and response latency from 1500 ms to 400 ms. • Identified rebate opportunities worth up to 5% of contract value within selected SKU subcategories through procurement dashboards analyzing pricing and revenue.
Senior Data Scientist
MY.GAMES
A video game developer and publisher with a portfolio of mobile and PC games, including titles supported by live operations and in-game advertising.
May 2023 — Dec 2025Improved game revenue forecasts and experiment planning, and used crash analysis to help prioritize fixes linked to organic install growth.
• Improved revenue prediction accuracy by 10% across 30% of advertising campaigns through cohort segmentation and validation. • Adapted long-term revenue models for mature cohorts, reducing forecast error by 10% for cohorts accounting for around 20% of total revenue, supporting portfolio planning and payback assessment. • Shortened experiment duration by 20% through a planning framework that estimated sample sizes and test duration using historical simulations and A/A checks.
Data Scientist
X5 Group
A major Russian food retailer operating nationwide grocery chains including Pyaterochka, Perekrestok, and Chizhik, alongside other retail formats.
Aug 2022 — May 2023Applied uplift targeting and A/B testing to retail campaigns, increasing conversions and selected campaign sales, and automated recurring analysis.
• Increased conversions per contact by 40% through CatBoost uplift targeting, coordinating campaign audiences under contact and product-category restrictions. • Optimized promotional campaigns through targeting insights and A/B testing, contributing to 10% higher sales for selected client campaigns. • Automated 30% of recurring data analysis tasks with PySpark.
Machine Learning Engineer
ANDATA
Sep 2021 — Aug 2022Built advertising recommendation and anomaly-detection tools, reducing keyword selection time, campaign workload, conversion cost, and false alarms.
• Built an advertising keyword recommender using NLP embeddings and a fine-tuned RuBERT ranker, generating daily batch recommendations that cut keyword-selection time by 80% and the campaign optimization team’s overall workload by 20%. • Lowered cost per conversion by up to 10% through automated audience segmentation while maintaining conversion volume. • Implemented an anomaly-detection microservice for marketing campaigns, reducing the false-alarm rate from 20% to 5%.
Data Scientist
Freelance
Jan 2020 — Sep 2021Freelance data science work included a news clustering tool and a customer-review sentiment analysis service.
• Developed and deployed a news clustering tool with word clouds for each subtopic. • Created a sentiment analysis service to identify issues and track trends in customer reviews.
