Yuri Dementiy

Head of Data Science at Wildberries. I drive security, Trust & Safety, and business-critical AI initiatives across ads, payments/banking, logistics, warehouses, and the core e-commerce platform. I lead ~80 specialists across 12 teams, with end-to-end ownership of risk, fraud, abuse prevention, and revenue-impacting ML systems.

I owned the Risky Goods handling stream in Trust & Safety, with end-to-end responsibility for detection, enforcement, and business results. In parallel, I contributed to broader antifraud and moderation initiatives, helping reduce platform risk and operational costs.

In my multifaceted role, I combined the responsibilities of CTO, CPO (for the division's products) and Head of Division (managing people), leveraging comprehensive tech expertise for our products. I was the driving force behind the launching new products and charting new directions. Under my leadership, we embarked on a journey that culminated in achieving State-of-the-Art products' quality and bringing the products to market. Our accomplishments include: - Forming and nurturing a cross-functional team. Together, we made several AI products, notably a service for anomaly detection in energy systems and AutoML solutions tailored for the Power Engineering sector. - Pioneering a new business direction grounded on the products we developed. - Introducing a new SOTA method that dramatically decreased the identification error rate from 3% to a mere 1%, all while ensuring the method remained production-ready. - Our AutoML service was a game-changer, cutting down the time-to-market for our new models by 30%. This efficiency did not compromise quality, which remained at the SOTA level. - I'm proud to have bolstered our team's resilience by increasing the bus factor by 2. - Achieving fundraising: I played a pivotal role in securing external financing for our development, which covered 100% of development expenses.

University Lecturer in Power Engineering Passionate about bridging the gap between traditional power engineering and the latest advancements in machine learning and deep learning. Developed and spearheaded a comprehensive ML/DL curriculum tailored for master's degree students in power engineering. Key Achievements: - Designed and implemented a ML/DL curriculum for power engineering master's students, ensuring they are at the forefront of industry innovations. - Mentored and guided numerous students on their research projects, fostering a culture of innovation and critical thinking. - Organized workshops, seminars, and guest lectures to provide students with diverse perspectives and hands-on experience in the field.

As the tech leader of a team of Data Scientists, I spearheaded the development of algorithms and ML pipelines, ensuring cutting-edge solutions and efficient implementations. - Built a UGC NSFW video moderation system, leveraging advanced ML techniques to ensure content safety and compliance. - Created a recommendation system for a wholesale business, achieving a remarkable 58% precision (precision@20 = 0.58) for 1000+ products. Notably, half of the recommended products had never been offered to customers prior to the system's implementation. Stack: Python, PyTorch, LightFM, Sklearn, SQL, Redis, Flask, LightGBM

Research and development of control algorithms Stack: Python, Matlab Developed and researched an algorithm for controlling the device, which became the basis for the company’s new product. The device successfully passed pilot operation.