Senior Machine Learning Engineer
Gazprombank
Oct 2021 — PresentBuild production LLM, RAG, agentic AI, NLP, personalization, and retail modeling systems; lead a cross-functional team and monitor 60+ models in production.
ER
Eugene Romanov
@wallrich
Senior ML/NLP Engineer | LLMs, RAG, AI Agents & MLOps
Senior ML/NLP Engineer with 6+ years of experience building production AI systems for banking and fintech.
Open to
Research collaborationProject collaborationStartup collaborationCompany partnerships
About
I specialize in LLMs, RAG, AI agents, PEFT fine-tuning, classical machine learning, and MLOps. I have led cross-functional teams and delivered scalable solutions for intelligent search, compliance, fraud detection, personalization, and business process automation with measurable impact.
Career
Experience: 6 yearsExperience
Middle Machine Learning Engineer
Alfa-Bank
Nov 2020 — Oct 2021Developed anti-fraud, NER, and adverse media screening systems for compliance, automating KYC workflows and reducing manual data entry.
Junior Machine Learning Engineer
QIWI Ltd
Sep 2019 — Nov 2020Built B2B credit scoring and hybrid NLP classification systems using CatBoost, neural networks, and BERT, improving risk decisions and feedback classification.
Education
Master of Science in Artificial Intelligence
ITMO University
2024 — 2026Master's program in Artificial Intelligence.
Bachelor of Science in Risk Analysis & Economic Security
Financial University
2015 — 2019Bachelor's degree in Risk Analysis and Economic Security.
Selected projects
AI Agent for Credit Application Analysis
Built a ReAct-based agent with Qwen3-14B-Instruct, PEFT LoRA, schema-aware tool-call validation, and execution trajectory control, reducing erroneous action chains by 15%.
RAG for Legal and Compliance Search
Designed a RAG system with Qwen2.5-14B-Instruct, fine-tuned BGE-M3 and BGE-Reranker, HyDE, and RRF, improving Recall@10 from 0.82 to 0.92.
Achievements
2.5× Increase in Campaign Response Rates
Gazprombank
Delivered a full-cycle retail modeling suite for churn, next-best-offer, uplift, and reactivation using CatBoost, TabNet, and RNN models.
Skills
PythonSQLPyTorchCatBoostLLMsRAGAI AgentsNLPMLOpsDocker