Apr 2024 – Present · 2 yrs 1 mo · United States · Remote

• Cooking large diffusion models for fashion e-commerce, providing unlimited, diverse, high-quality visuals to transform product photography and streamline content creation.

Feb 2023 – Sep 2024 · 1 yr 7 mos · Dubai, United Arab Emirates · On-site

• I've built and led a team at an AdTech startup, focusing on machine learning initiatives across the Demand Side Platform (DSP), Data Management Platform (DMP), and Antifraud Platform.
• Key Contributions:
• 🚀 Fraud Detection: Spearheaded the implementation of an anomaly detection model achieving 20% reduction of costs, <2 ms inference time, handling up to 150k requests per second seamlessly.
• 🚀 Pioneering Graph Neural Networks: Engineered Large Graph Neural Networks (> 3B parameters) for obtaining user and application embeddings, enhancing high-level models like look-alike, targeting and etc.
• 🚀 Advanced Targeting Solutions: Researched and developed targeting solutions to optimize click through rate (CTR), installs per mile (IPM) and other types of conversions.
• 🚀 Integration of ML models into High-Load DSP: Led the research of integrating ML models into our high-load DSP environment (> 150k requests per second), guaranteeing bid response times < 50 ms. We achieved reduced inference time by optimizing feature storage using Aerospike for fast access to embeddings, leveraging the C++ API for PyTorch, and optimizing communication between ML applications and DSP.

Mar 2022 – Feb 2023 · 11 mos · Limassol, Cyprus · On-site

• At Exness Data Office, I led ML projects focusing on detecting trading fraud. I was responsible for the entire project lifecycle, collaboration between an engineering team and stakeholders, technical design, roadmap creation and execution.
• Key Contributions:
• 🚀 Enhanced Analyst Efficiency: Developed a real-time service that reduced False Positive alerts for Anti-Fraud analysts by 40%.
• 🚀 NeurIPS Competition Winner: Secured victory in the Trojan Detection Challenge at NeurIPS, the world's premier ML conference. Achieved exceptional results with a 99.98 AUROC in classifying hacked Deep Neural Networks by leveraging a combination of backdoor scanning methods and network weight analysis.

Apr 2021 – Mar 2022 · 11 mos · Moscow, Russia

• Led a dedicated Data Science team at Kaspersky, evolving from a Data Scientist in the Anti-Spam department to a Group Manager overseeing ML initiatives across the company. My focus was on enhancing Kaspersky's security solutions through innovative Machine Learning technologies.
• Key Responsibilities:
• 📍 Managed a 14-member ML team, driving strategies to elevate detection quality for Kaspersky's product suite.
• 📍 Orchestrated 9 ML projects and supervised 3 external research endeavors, resulting in the development of four patented technologies integrated into flagship Kaspersky products.
• 📍 Oversaw team strategy, streamlined processes, and fostered a culture of innovation.
• 📍 Collaborated with stakeholders, aligning efforts to achieve optimal outcomes in projects and initiatives.
• 📍 Led hiring initiatives, ensuring the recruitment of top-tier talent to drive Kaspersky's data-driven initiatives forward.
• Key Contributions:
• 🚀 DeepQuarantine: Deployed a Deep Convolutional Neural Network to improve email protection components, reducing False Negatives by 30%.
• 🚀 Kaspersky MDR Autoanalyst: Engineered a crucial component of Kaspersky Managed Detection and Response, significantly reducing False Positive alerts by 50% and contributing significantly to the product's success in 2021.
• 🚀 Anti-Spam Autoanalyst: Automated the processing of 80% of incoming spam messages for products like Kaspersky Secure Mail Gateway, revolutionizing Anti-Spam analyst workflows.

Apr 2020 – Apr 2021 · 1 yr · Moscow, Russia

Sep 2018 – Mar 2020 · 1 yr 6 mos · Moscow, Russia

Sep 2017 – Sep 2018 · 1 yr · Moscow, Russia