Ivan Cherdantsev
Ivan Cherdantsev

AI Software Engineer

Open to offers · Member since 16 Nov 2026
Location
Porto, Portugal
Desired salary
Unspecified
Work preference
Remote Only / Full Time, Contract
Experience level
Lead

About

Professional summary

I am an AI Software Engineer and data science leader with 11 years of experience solving complex, poorly formalized business problems from early concept through production deployment.

My recent focus is on production LLM agents, retrieval-augmented systems, and secure enterprise AI platforms. I build practical solutions using technologies such as AWS Bedrock, Claude, LangChain, smolagents, Qdrant, Python, and FastAPI.

I have delivered data products across clickstream analytics, fraud detection, personalization, banking segmentation, industrial optimization, and forecasting. My work has generated measurable impact, including lower annotation costs, improved model precision, multimillion-dollar business opportunities, and scalable data pipelines.

I combine hands-on engineering with machine learning expertise, including PyTorch, CatBoost, LightGBM, scikit-learn, PySpark, and SQL. I am experienced in building and operating cloud-based ML systems on AWS using SageMaker, Athena, Docker, and production data workflows.

I am also an experienced technical leader. I have hired, mentored, and led cross-functional teams, developed team leads, managed stakeholder relationships up to C-level executives, and taken ambiguous client opportunities from pre-sales through delivery.

I am based in Porto, Portugal, and work remotely. I communicate professionally in English and am a native Russian speaker.

Skills

31 capabilities

Tech stack & tools

Working toolkit

Development

Languages & Frameworks

Experience

Career history

Head of Data Science Datos (an Adobe company)

Led data science for a clickstream-data-as-a-service business remotely from Portugal. Built a production, tool-using LLM enrichment agent using Claude on AWS Bedrock that processes approximately 50,000 records monthly, supports human review for low-confidence cases, covers more than 3,000 active tickers and 87% of events, and reduced annotation costs by about 80%.

Hardened the system with hybrid retrieval using semantic search, BM25, reranking, Qdrant, and human-verified fact memory. Designed a statistical PII detector in PySpark, scaled data pipelines processing hundreds of GB per day, delivered a 95%-accurate SageMaker language detector that enabled an $11M annual deal, and led a cross-functional team of four.

Founding Engineer Corporate LLM Chat

Built an internal enterprise LLM chat platform end-to-end as the sole engineer, covering data strategy, tooling, implementation, validation, and live client demos. Replaced a legally flagged core dependency without disrupting the database layer and supported two external client pilots.

Designed storage-level tenant isolation through dedicated Qdrant collections and per-tenant buckets with opt-in sharing. Built a smolagents analytics agent that generates reproducible code in an isolated sandbox and implemented Russian-language guardrails for prompt injection and toxicity.

Machine Learning Engineer Playrix

Developed an offer-personalization pipeline for a major mobile game publisher, covering ETL, machine learning modeling, and offline metric evaluation.

Used a PySpark recommendation-system framework for offline validation, with estimated results indicating a 15% ARPU lift.

Team Lead Data Science Rubbles

Led data science consulting engagements across FMCG, logistics, heavy industry, and banking. Improved rotor-balancing optimization at a nuclear-industry plant, reducing iterations from five to three and generating approximately $700K in annual value; also delivered railway-freight forecasting that supported nearly $1M in margin.

Built and deployed a banking user-segmentation service used by four top-10 banks and serving more than 100 million users daily. Converted three of eight pre-sales opportunities into delivery projects, grew the team from four to 15 people across three domains, and developed three internal team leads.

Senior Data Scientist Freelance / Contract

Worked as a senior data scientist on contract engagements, including as part of the data science team at Cybertonica, a transaction anti-fraud startup.

Developed a fraud-detection system and engineered new features that improved the precision of existing models by 13%.

Education

Learning history

South Ural State University (National Research University)

MSc, Applied Mathematics

This professional hasn’t added portfolio projects yet.

This professional hasn’t listed any services yet.

Jobs Talent AI Tools Salaries
Menu