Download PDF

Ivan Reshetnikov

Senior ML / AI Engineer

Belgrade, Serbia | resivalex@gmail.com | linkedin.com/in/resivalex | github.com/resivalex

Professional Summary

Senior ML / AI Engineer with 10+ years in software engineering, including 5+ years in applied machine learning.

I build production ML systems for real business problems: matching, ranking, duplicate detection, pricing, anomaly detection, forecasting, and LLM/RAG automation. I usually work end-to-end, from understanding the problem and defining metrics to building, deploying, and improving the system after release.

Work Experience

Senior ML Engineer

Maksmart

Aug 2024 - Apr 2026

Built production ML systems for a B2B marketplace: product matching, duplicate detection, pricing, anomaly detection, catalog quality, delivery forecasting, and LLM/RAG-assisted data workflows.

  • Shipped 8 ML models to production across matching, pricing, catalog quality, and forecasting.
  • Built a product matching system that reduced unmatched supplier offers by 2x.
  • Developed duplicate scoring that reduced implicit product duplicates by 60% at 95% precision.
  • Built price regression with uncertainty estimates to support supplier offer moderation.
  • Implemented price anomaly detection to identify unit-of-measure issues, data-entry errors, and suspicious supplier prices.
  • Implemented last-mile delivery time forecasting with under 15% error.
  • Automated type definition, merge decisions, and data labeling with an LLM/RAG workflow, supporting product type discovery across the majority of the catalog.
  • Optimized data processing and data quality pipelines, reducing memory consumption by 5x.

Data Scientist

Architech NYC

Apr 2022 - Jul 2024

Built production ML, analytics, and automation tools for transportation and business operations: rate calculation, document validation, BI reporting, and email workflow automation.

  • Built a container transportation Rate Calculator with 11% RMSPE, monitoring, and regular retraining.
  • Developed an OCR-based document validation service with 95% precision, reducing dispatcher workload.
  • Built BI dashboards for transportation operations, improving visibility and reducing manual reporting.
  • Automated email handling with the ChatGPT API, answering status requests and drafting rate quotes for dispatcher approval.
  • Investigated data consistency issues and prepared ad-hoc analyses for current business needs.

Machine Learning Engineer

Sape

Dec 2020 - Mar 2022

Applied ML to SEO and search products operating on millions of documents: semantic search, topic classification, text similarity, and price recommendation.

  • Built a semantic backlink search system across millions of documents, achieving a 4x speedup through ClickHouse sharding.
  • Replaced mostly tag-based or random backlink discovery with similarity-based recommendations.
  • Improved backlink suggestion relevance, contributing to positive user feedback and continued investment in the feature.
  • Developed topic classification, text similarity, and price recommendation models for SEO workflows.
  • Analyzed product usage and dissatisfaction signals to support SEO specialists and product decisions.

Lead Ruby on Rails developer

Sape

Dec 2016 - Nov 2020

Led backend and full-stack development for an ad-tech product integrating multiple advertising platforms. Worked on technical decisions, API design, code review, testing, deployment, monitoring, and frontend modernization.

  • Unified API usage for external users and the frontend, reducing duplicated integration logic.
  • Built tools for advertising analytics, moderation-state synchronization, and campaign management.
  • Added system problem-detection tooling to make operational issues easier to find and investigate.
  • Implemented interactive UI features such as weekly schedules, bulk copying, and ad previews.
  • Supported frontend migration from CoffeeScript/jQuery to ES6 and React.
  • Helped unify account-wide UI layout and styling.

Ruby on Rails developer

ORDIN

Apr 2016 - Nov 2016

Supported and developed Ruby on Rails applications for real estate process management and professional qualification workflows, across backend features, tests, deployment, and legacy frontend maintenance.

Software developer

Nasca LTD

Apr 2015 - Jan 2016

Web and trading-related development in a Scrum team: a public website built from scratch, and C++ components of a multi-part network application for exchange workflows, including exchange interaction and the Qt client.

C++/Qt developer

Nasca LTD

Dec 2012 - Sep 2013

Trading-related desktop and backend tools: trading engine improvements, exchange log parsing, transaction logging, and an exchange emulator for testing and development.

Certifications

Claude Certified Architect – Foundations by Anthropic

Aug 2026 - Aug 2027

Agentic architecture and orchestration, Claude Code workflows, prompt engineering and structured output, tool and MCP integration, context management and reliability.

Core Skills

Machine Learning & Modeling: Python, statistics, machine learning algorithms, PyTorch, scikit-learn, LightGBM, gradient boosting, feature engineering, hyperparameter tuning, regression, classification, clustering, neural networks, model evaluation, Natural Language Processing (NLP), OCR, embeddings, fastText, sentence-transformers, metric learning, active learning, recommender systems, AutoML, Optuna

Production ML & MLOps: MLOps, model deployment, model monitoring, scheduled retraining, model versioning, guardrail metrics, CI/CD, GitLab CI, automated testing, pytest, Docker, FastAPI, Google Cloud Platform (GCP), PostgreSQL, ClickHouse, SQL

Search, Matching & Ranking: product matching, duplicate detection, e-commerce search/catalog, information retrieval, semantic search, similarity search, vector search, vector database, candidate generation, approximate nearest neighbor search, reranking, fuzzy matching, FAISS, Elasticsearch

LLM, Agents & RAG: Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), AI agents, agentic architecture, agent orchestration, multi-agent systems, prompt engineering, structured output, structured extraction, tool calling, Model Context Protocol (MCP), context management, Claude API, Claude Code, Claude Agent SDK, OpenAI API, Google Vertex AI, LLM evaluation, automated data labeling, human-in-the-loop workflows

Anomaly Detection, Pricing & Forecasting: anomaly detection, price anomaly detection, price regression, uncertainty estimation, confidence scoring, calibration, quantile regression, forecasting, delivery time forecasting

Data Engineering & Analytics: Pandas, NumPy, Parquet, scheduled pipelines, Celery, sharding, Apache Superset, data quality analysis

Software Engineering & Observability: programming, backend engineering, Python application architecture, API design, REST APIs, system integration, Pydantic, mypy, code review, Sentry, Grafana

Education

Pskov State University
Master's degree, Informatics (Information Systems and Technologies)
2008 - 2013

Languages

English (Advanced), Russian (Native)