en ru

CV

AI Architect & Founding ML Engineer

I bridge the gap between theoretical machine learning and physical industrial production. With over 10 years of experience in data-driven development, hardware engineering, and applied chemistry, I focus on building autonomous, closed-loop systems that operate reliably in harsh, real-world conditions.

Currently, I operate in two complementary domains. As an AI Architect, I design enterprise-level RAG and LLM orchestration pipelines, strictly eliminating AI hallucinations for high-stakes banking and construction sectors. Concurrently, as the Founding ML Engineer at LogicYield, I architect CPU-optimized Edge AI sensors that digitize complex inert processes, saving industrial plants millions of dollars annually in raw materials and defective products.

Key Skills:

Edge AI & Computer Vision 100%
LLM Orchestration & RAG 90%
Industrial Automation & IoT 90%
Python & PyTorch 100%
MLOps & Infrastructure 85%
Team Leadership 100%

Professional Experience

Founding ML Engineer (2022 - present) LogicYield LLC (Startup)

Specializes in intelligent solutions for heavy chemical industries, reducing $1.8M of annual losses per plant. Resident of Science innovation center.

Responsibilities:
  1. Architected autonomous Edge AI optical sensors for the chemical and fertilizer industries, processing complex visual streams without cloud dependencies or heavy GPU infrastructure.
  2. Setting up MLOps processes and managing AI models for deploying and optimizing machine learning systems.
  3. Develop R&D projects in AI and quality control in chemical industry.
  4. Managing team (5 employees).
  5. Developing ML and CI/CD pipelines and tests.
  6. Presenting solutions to industry leaders and enterprises.
  7. Preparing the research and development budget.
Achievements:
  1. Developed DotPulse devices, a real-time granulometry system operating directly on active conveyor belts, reducing factory reaction speed to deviations from 4 hours down to 5 minutes, eliminating an estimated $1.8M in annual defect losses per plant.
  2. Built GuardDetector devices, a CPU-optimized temporal sequence-tagging system (MobileNetV3 + BERT) for occupational safety and maintenance tracking, ensuring privacy by design while preventing operational bleed.
  3. Attracted $150,000 of investments.
AI Architect (2025 - present) Digital Biz Factory LLC, remote

Company specializing in intelligence IT development. Resident of IT innovation center. Responsibilities:
  1. Design and deploy enterprise RAG and AI systems tailored for banking and construction sectors, enhancing data retrieval efficiency (Text Search Systems Design, Domain-Specific Text Embeddings, Text Classification, Text Clustering, LLM Domain Adaptation).
  2. Setting up MLOps processes and managing AI models for deploying and optimizing machine learning systems.
  3. Develop production-ready software engineering pipelines incorporating Named Entity Recognition/NER, Knowledge Graphs, Vector databases and BERT.
  4. Construct AI Agents: Full-cycle AI Agent development and Structured Output parsing.
  5. Presenting solutions to C-level managers and users.
Achievements:
  1. Engineered 2 text search systems and domain-specific knowledge bases, prioritizing strict data retrieval architectures to eliminate model hallucinations and ensure absolute data reliability.
  2. Implemented structured NER extraction engines to map unstructured enterprise data into queryable formats.
Head of Machine Learning (R&D) (2022 - 2025) Autodoria LLC, remote

Specializes in intelligent transport solutions, reducing road deaths by 51% and accidents by 15.6%

Responsibilities:
  1. Management of the development and training of machine learning models and data processing for computer vision projects.
  2. Setting up MLOps processes and managing Torch models for deploying and optimizing machine learning systems.
  3. Supporting R&D projects, developing and implementing research and development strategies in the field of computer vision and scene analysis.
  4. Management of the cross-functional team of 12 ML engineers across multiple time zones, developing intelligent transportation and computer vision systems.
  5. Developing ML pipelines, automating, and testing model training processes.
  6. Overseeing the lifecycle management of models, evaluating market and scientific trends.
  7. Presenting solutions to senior management and clients.
  8. Preparing the research and development budget.
Achievements:
  1. Managed a cross-functional team of 12 ML engineers across multiple time zones, developing intelligent transportation and computer vision systems.
  2. Established comprehensive MLOps infrastructure (ClearML, GitLab CI/CD, Pytest), increasing team efficiency and raising code readability metrics from 3.0 to 9.2 (via pylint).
  3. Introduced code review and live coding practices (weekly sessions), boosting the development team’s efficiency from 1 to 5 merge requests per month.
Machine Learning Engineer (2021 – 2022) rebels.ai, remote

Company specializing in AI and software integration

Responsibilities:
  1. Setting up and maintaining the infrastructure of 2 servers with 4 GPUs.
  2. Data preparation (climate data, satellite data: sentinel, Google Earth Engine, Nasa, ground station data: CO2, PM 2.5, PM 2.5, PM 10, industrial data).
  3. Preparation of 20+ jupyter notebook (pandas, geopandas, matplotlib, sentinelhub, gee, numpy).
Achievements:
  1. Configured and maintained server infrastructure for data preparation and neural network model building.
  2. Developed 3 neural network models and actively researched neural networks.
  3. Analyzed and worked with data of up to 500 GB, optimized computations – reduced memory usage from 54 GB to 13 GB, implemented projects in climate and satellite monitoring.
Head of Automation Department <- Senior <- Junior Researcher (2011 - 2022) PhosAgro inc., JSC "NIUIF"

Top-tier research institute specializing in phosphate rock processing technologies

Responsibilities:
  1. Led a team of 4 engineers at Russia's leading phosphate rock processing research institute, bridging chemistry, engineering, and automation.
  2. Managing laboratory infrastructure.
  3. Performing statistical analyses of data and planning experiments: development of new techniques and methods, data collection and accumulation (SCADA, MES).
Achievements:
  1. Created a team from scratch: searched for funding, selected equipment, hired staff.
  2. Developed and launched 2 types of industrial control devices for physical and chemical analysis.
  3. Directed strategic R&D projects, successfully increasing departmental funding from $174,000 to $294,000.
  4. Authored over 30 industrial reports and modified technology schemes across multiple production facilities to integrate new automated processes.
Associate Professor, Senior Researcher (2018 - 2023) Cherepovets State University, Department of Automation and Control and Department of Chemical Technologies

Leading research institute in the field of mineral fertilizer production.
  1. Managed laboratory infrastructure; performing lectures and researches in ESG, industry automation systems and projects planning.
  2. Published 2 methodological manuals and 5+ scientific papers in Q2-Q3.

Patents, Key Publications & Intellectual Property

Key Scientific Publications
Patents & Registrations

Open-Source Code & Architecture Examples

My complete open-source portfolio is available on GitHub (DimYun). Below are key examples demonstrating my approach to MLOps and production-ready AI services.

Satellite Computer Vision: Multi-Label Classification & API
A multi-label classification neural network solution for Amazon forest satellite imagery. The primary focus of this project is demonstrating "industrial-grade" MLOps and code architecture, rather than just model accuracy.

Core Stack & Engineering Practices: PyTorch Lightning, timm, ClearML, strict linters (black, isort, nbstripout, flake8), data validation with Pydantic, and DVC for data version control.

Disclaimer: The repository contains short-trained versions of the neural networks (approx. 15 epochs) optimized for architectural demonstration.

Skills: Computer Vision · Machine Learning · MLOps · FastAPI · PyTorch · Neural Networks

End-to-End OCR Pipeline: Car Plates | FastAPI | TON
A complete, deployable service for vehicle license plate recognition, demonstrating the full lifecycle from edge-case segmentation to user-facing API deployment.

Project Architecture:
  1. Instance Segmentation: Neural network model to isolate the specific car plate area and filter noise.
  2. Optical Character Recognition (OCR): Sequence-recognition model for reading plate characters.
  3. Backend Orchestration: A FastAPI service integrating both neural networks into a seamless pipeline.
  4. Web3 & Bot Integration: A Telegram bot interface utilizing the API, featuring a TON blockchain payment gateway (Developed during the TON x ETH Belgrade hackathon).

Skills: Computer Vision · FastAPI · PyTorch · API Architecture · Web3 Integration

Education

Recommendations

Timur Fatykhov | my teacher at CVRocket course
I had the pleasure of teaching Dmitrii Iunovidov in my Machine Learning (CVRocket) course at DeepSchool. Throughout our time together, I was consistently impressed by his commitment to learning, enthusiasm, and deep passion for knowledge.

As a student, Dmitrii consistently demonstrated a high level of eagerness and intellectual curiosity, actively engaging with course material. He actively participated in class discussions, asked insightful questions, and consistently produced exemplary work. This earned him a place among the top 3 graduates of the sixth cohort of the course.

Furthermore, Dmitrii brought and developed his own project for car plate OCR, which was highly regarded by my colleagues.

What truly stood out about Dmitrii was his ability to think critically and analyze complex issues. He possesses a natural aptitude for problem-solving, and his analytical skills are sure to serve him well in his future endeavors.

Marat Gilmanov | managed me directly in Avtodoria LLC
I have been working with Dmitrii Iunovidov, the Head of Machine Learning at Avtodoria LLC, since 2022. Throughout our collaboration, I have been consistently impressed by his innovative approach, strategic thinking, and expertise in developing ML content and strategy.

As the Head of Machine Learning (R&D), Dmitrii has a deep understanding of our target audience's needs and the ability to create compelling content that resonates with them. He has a natural talent for finding unique perspectives and approaches that set our content apart from the competition. Additionally, he has developed a robust and effective R&D infrastructure using Ubuntu Server, CVAT, ClearML, Nginx, Docker, and FastAPI.

Dmitrii's strategic thinking and ability to develop effective R&D content campaigns have been particularly impressive. He has a keen eye for what works and what doesn't, and has been instrumental in helping our company achieve our R&D goals.

Not only is Dmitrii skilled in his technical abilities, but he is also a pleasure to work with. He is a strong communicator and always goes the extra mile to ensure the success of our campaigns.

I highly recommend Dmitrii Iunovidov for any opportunity in the field of R&D content development and strategy. He has consistently demonstrated his ability to excel in his role. It has been a pleasure to work with him, and I am confident that he will be a valuable asset to any team or organization.

Ikechi Ndukwe | my colleague at LogicYield LLC
Dmitrii is an exceptional leader and mentor in the fields of Machine Learning, Deep Learning, and Computer Vision, applying these technologies to solve real-world problems. He is also a highly skilled MLOps Engineer. Beyond his technical expertise, Dmitrii is an experienced researcher and a dedicated team player who is always eager to support his colleagues and mentees. His guidance has been immensely beneficial to my PhD research, providing valuable insights and direction.

Working with Dmitrii has been a great pleasure, and I wholeheartedly endorse him for any position in the fields of Machine Learning and Research & Development.

Courses

Languages

  • English (C1, proficient)
  • Russian (native)
  • Bulgarian (B1, intermediate)

Other

Awards
  1. Winner of the Math & Physics Olympiad (National Level)
  2. "Engineer of the Year" - National Science and Engineering Association, 2022
  3. 1st Place, "Innovative Engineer of the Year" - Regional Government Award, 2021
  4. 1st Place in corporate "Young Leader" competition - Top-tier Chemical Corporation, 2019
  5. Certificate of Appreciation - National Academy of Sciences, 2019
  6. Honorary Diploma - National Union of Chemists, 2019
About me
From flasks to code: I have navigated a thrilling career journey, transforming from an analytical chemist into a founder and AI Architect. Driven by a passion for science and a belief in building things that actually work in the physical world, I leverage my engineering PhD, Python, and PyTorch to solve complex industrial challenges. A Linux fan at heart, I feel a little uncomfortable without my computer nearby.

Outside of the lab and the terminal, I hold ISSA Inshore Skipper, boat master, and car licenses—and I love to use them all!