Machine Learning Engineer · Applied AI Technical Lead

Every research claim below is backed by an independent, externally verifiable source — not just text.

I'm an ML Engineer and Applied AI Technical Lead with 5+ years of experience building RAG and document-intelligence platforms, medical imaging systems, and scientific ML tools. I lead teams of 3–5 engineers end to end — architecture, delivery, code and solution review, and ML engineering quality — and I still write the hard code myself: tensor methods, Bayesian inference, segmentation, radiomics, and production Python/FastAPI systems that run in CI/CD, not just in notebooks.

If you're evaluating this profile as a recruiter, collaborator, or hiring manager: the short version is — I ship research-grade ML into production, I lead the team that ships it, and I publish/patent what we learn along the way.

What I bring to a team:

• 🧭 Technical leadership — currently leading architecture and delivery of three applied AI platforms simultaneously, coordinating 3–5 engineers per team, owning system boundaries and operational readiness.
• 🔬 Research-to-production range — equally comfortable deriving a Bayesian sensor-placement algorithm and shipping the FastAPI service, Redis workers, and CI gates that put it in front of users.
• 📈 Measurable impact — e.g. cut CT/MRI segmentation processing time by 40% with an automated segmentation + manual ROI-correction pipeline.
• 🏛️ Recognized track record — 5+ publications (incl. arXiv), one officially registered software product, oral presentations at international conferences (Singapore, China), and multiple AI/ML competition wins.

• Machine Learning Engineer / Technical Lead, Analytics and Machine Learning Department, MSUU (Apr 2024 – Present) — leading architecture and delivery of three applied AI platforms (CourseLLM, Modular RAG Platform, LLM Adviser — case study), coordinating teams of 3–5 engineers.
• Data Scientist / Research ML Engineer, Heriot-Watt TPU Center (Sep 2024 – Present) — tensor-based reduced-order modeling and QR sparse-sensor placement for reservoir data; published NAB as a tested, benchmarked Python package, and the underlying methods as public research code + arXiv preprint.
• Machine Learning Engineer, Cardiology Research Institute (Mar 2021 – May 2025) — built EPIFAT, officially registered cardiac CT segmentation software (No. 2025610317); cut processing time by 40% — related public research code: radiomics.

Full role-by-role breakdown, day-to-day scope, and team details — happy to walk through in a call, or see LinkedIn once linked below.

Applied AI platforms I lead in production (institutional/client work — repos are private, happy to walk through architecture and results in a call):

Public research code — the methods behind the résumé, actually on GitHub:

• 🔗 tensor_based_modal_decomposition_method — Tucker/HOSVD-based modal decomposition and QR sparse-sensor placement for reservoir field reconstruction; the working code behind the AI4X Singapore 2026 talk and the arXiv preprint.
• 🔗 radiomics — radiomic texture analysis of cardiac polar maps with ML, the research line that led into EPIFAT.

• "Tensor-Based Modal Decomposition and Sparse Sensor Placement for the Brugge Field Simulation Model." arXiv preprint, 2026.
• "Approach to Identifying Key Areas for Further Reservoir Study Using Tensor-Based Modal Decomposition." Conference paper, 2025.
• "Automatic Segmentation of Epicardial Fat and Quantitative Evaluation of Radiomic Parameters in Cardiac CT." Proc. XXI Int'l Conf. "Perspectives of Fundamental Sciences Development," 2024.
• "Capabilities of Radiomic Analysis of Cardiac MRI Images in Cine Mode for Identifying Post-Infarction Areas." Digital Diagnostics.
• "Beam Parameters Restoration at the NICA Accelerator Complex Based on the Beam Position Monitor Data." Proc. START, JINR, 2023.

• 🎤 Oral Presentation — "Tensor-Based Modal Decomposition with QR Pivoting for Sparse Sensor Placement and Field Reconstruction," AI4X – Accelerate Conference, Singapore, 2026.
• 🎤 Accepted Presentation — "Probabilistic Evaluation of Parameter Space Using Neighbourhood Algorithm Bayes," Data Intelligence in the Oil and Gas Industry, Nizhny Novgorod, 2026.
• 🌏 Participant, School-Conference on Tensor Methods in Mathematics and AI, Shenzhen, China, 2024.
• 🌏 Participant, Skoltech-HIT Summer School, Harbin, China (remote), 2025.
• 🥇 Winner, FINOdays AI/ML Track — 🏅 Special Nomination, National Technology Olympiad — 🏅 Prize Winner, MIPT AI/Math/Physics Hackathon.
• 🥈 Second-Degree Diploma — Cardiac CT segmentation & radiomics presentation, Perspectives of Fundamental Sciences Development, 2024.
• 📜 Software No. 2025610317 — official state registration for EPIFAT (issued Jan 9, 2025).

Tomsk Polytechnic University

• M.S., Applied Mathematics and Computer Science — 2024–2026
• B.S., Applied Mathematics and Computer Science — 2020–2024
• Diploma of Professional Retraining, Data Science and Machine Learning — 2023–2024

Skoltech (Skolkovo Institute of Science and Technology)

• Professional Development — Large Language Model-Based Agents, 2025
• Professional Development — Generative Models Based on Adversarial Learning, 2024

I'm always open to talking about RAG/LLM systems, medical imaging, scientific ML, and technical leadership roles — whether that's a full-time role, research collaboration, or consulting.

CV_SamatovDS.pdf is © Denis Samatov, provided for reference in hiring/collaboration contexts only — all rights reserved, no redistribution without permission.

⭐ Thanks for stopping by — if something here is relevant to what you're building, reach out.