Mikhail Borodastov

Driving ML product strategy across fraud prevention, credit risk, and AI-powered customer support at Uzbekistan's leading fintech ecosystem ($2.3B valuation, 20M+ MAU).

• Owned 20+ production AI services across text, image, video and multimodal domains (LLMs, VLMs, embeddings, vector search, RAG), used as reusable platform capabilities by multiple internal and external teams • Grew internal LLM/VLM platform to 900 users (DAU 300) • Led end-to-end delivery: discovery, requirements, technical trade-offs, production rollout, scaling and adoption across teams • Orchestrated cross-functional execution across engineering, MLOps, DS and business stakeholders, aligning on priorities, delivery plans and success metrics • Owned platform reliability and governance: usage limits (tokens, RPS), monitoring, incident management and postmortems to ensure predictable partner experience

As a PM at Uzum, I was engaged in developing and launching new products within the diverse business units of our digital services ecosystem. My involvement centered around integrating AI and ML technologies. • Contributed to the development of an ML-based Credit Scoring service from scratch, achieving a 30% improvement in model quality (measured by GINI), potentially allowing for a 14% increase in Approval Rate without impacting the Non-Performing Loan • Spearheaded the development of a MVP Lead Generation service, leveraging big data to drive effective advertising campaigns targeting over 30M+ users • Conducted extensive market research and interviews to optimize CJM and conversion rates for an MVP lead generation service • Monitored pilot launches by collecting and preprocessing row data from multiple databases with following analysis and visualization

OneFactor, recognized by Gartner as a 'Cool Vendor' for its secure data monetization platform based on AI & ML, is used by over 80% of Russian banks for risk assessment and lead generation. At the end of 2022, OneFactor was integrated into Uzum, Uzbekistan's #1 Digital Ecosystem, becoming a key component of its services. • Bring to the market a bank's Anti-Fraud service for 5 months, reaching $600K in the 1st year • Proposed a feature generation technique that increased the quality of Credit Scoring ML models by 3 pp GINI, compared to a 1 pp increase over the previous two years before my tenure • Developed and managed product roadmaps, prioritizing features (OKR, LEAN, RICE) and releases based on ROI estimations • Set goals and defined tasks for teams of data scientists and data engineers, fostering a collaborative and productive environment in Agile settings

GlowByte is an international group offering professional services in financial analytics, risk management, data warehousing, advanced analytics, and Industrial Internet of Things. • Built ML models that served as the foundation for digital advisors in manufacturing plant, leading to a 1% increase in cement productivity • Developed a regression model estimating the percentage of the target fraction at the output of the ore grinding mill, which formed the basis of a virtual grinding sensor used for real-time process monitoring • Conducted exploratory data analysis with report preparation in Jupyter Notebook and data visualization with Matplotlib for various technological processes, including ore beneficiation, gas processing, cement production, and paper manufacturing • Prepared data pipelines for gathering and processing data from diverse sources, and generating features for machine learning model development

•Managed sales of major semiconductor brands including Intel, Xilinx, Texas Instruments, NXP, and Microchip. •Achieved a $5 million revenue target by developing existing and new customers

Developed and implemented a new promotional strategy and educated the sales team, which effectively repositioned the brand and doubled revenue

Key study areas included electronics, sensor signal amplification, physics, advanced mathematics, and ionizing radiation registration techniques.