Aug 2024 – Jun 2025 · 10 mos · London Area, United Kingdom · Hybrid

• Vertical GenAI

Feb 2023 – Oct 2023 · 8 mos · Dubai, United Arab Emirates

• Business management system for pet beauty salons ($20B market!)
• Founded, hired and led the high-velocity engineering team of 5 members, successfully shipping the core product within 3 months by guiding the team through technical hurdles in robust authentication/authorization (Auth0 implementation) and multi-tenant SaaS setup
• Led development of the complex platform (infra, backend, user-facing web app, admin web app, user-facing cross-platform native app)
• Designed the end-to-end system architecture and set up cloud and code infrastructures (CI/CD, code generation, artifact publishing, app publishing with Fastlane)
• Balanced the need for rapid development strategically addressing architectural elegance versus speed, and refactoring based on product features.
• After my co-founder (on the commercial side) gave up I hacked marketing & sales. Developed and executed a scalable outbound sales pipeline, acquiring over 40 leads and attracting initial paying customers.
• Tech stack: AWS infra (extensive - compute, storage, event sourcing, scheduling, emails, etc), Auth0 for authentication, authorization, Python backend (FastAPI, alembic, SQL Alchemy), PostgreSQL operational db (+ row level security for multi-tenant set up), React for web app, Flutter for native apps

Sep 2021 – Dec 2022 · 1 yr 3 mos · Worldwide · Remote

• rexy.ai is a no-code platform for AI on satellite and drone images, empowering analysts to train computer vision models without coding.
• Finalist of the prestigious Gravity Innovation Challenge
• Finalist of the pitch competition at the International Astronautical Congress.
• Developed the core AI and backend components including serverless model training, active learning, data labeling, high-performance inference
• Developed a specialized object detection model to detect small objects on satellite images in real time. Optimized model to able to be served using ONNX Runtime and NVidia Triton inference server
• Hired and managed contractors to develop customer-facing web app
• Leveraged deep tech knowledge to pivot the product strategy from focusing solely on AI inference to building a UI-driven system for non-technical business users to train models, which unlocked several pilot projects.
• Successfully secured several pilot projects across the world (UAE, Japan), learned business development and client relationship building skills.
• Gained recognition from the UAE Space Agency and secured grants from AWS and Nvidia.

Jul 2021 – Sep 2021 · 2 mos · On-site

• Led team of 18 people, including software engineers, ML engineers, and domain experts, overseeing technical strategy and project execution for a complex project with AI/ML at its core.
• Project management, strategic product planning and complex stakeholder management
• Guided the team through significant technical challenges related to deployment in an air-gapped client environment and handling large multi-gigabyte images for browser rendering and AI processing, encouraging collaboration with DevOps and optimizing storage formats for efficient retrieval.
• Primarily provided strategic engineering oversight and client-side troubleshooting in this role.

Feb 2019 – Jul 2021 · 2 yrs 5 mos · On-site

• Established and led a high-performing team of 7 ML engineers and researchers.
• Achieved one of the highest team satisfaction rates within the organization, with zero voluntary team departures over a 2-year period, reflecting strong leadership and team management fostered through bi-weekly one-on-ones, empowering team ownership, and encouraging internal presentations.
• Managed contributions to a complex $23 million project involving satellite image processing and analysis, pushing the boundaries of technology in spacetech.
• Designed and oversaw the development of a highly scalable distributed platform for running AI/computer vision models on satellite images, making key technical decisions for a fully asynchronous system with specialized microservices, orchestrated by an in-house real-time scheduler, and efficiently managing GPU resources; personally developed core components including the scheduler and asynchronous logic.
• Personally developed a novel computer vision algorithm for real-time object detection and led the in-house development of this state-of-the-art model for satellite images, resulting in a paper submission to the top AI/computer vision conference (ECCV).
• Designed and implemented a robust process for continuous data acquisition, model training, and production deployment using in-house system for annotation delivery, initially facilitating collaboration and later delegating management.
• Built strong collaborative relationships with the G42 research lab (Inception Institute of Artificial Intelligence), effectively leveraging cutting-edge research for practical applications.
• Cultivated warm and trustworthy relations with the client, designing and running a series of educational workshops to enhance understanding and collaboration.

Jan 2018 – Feb 2019 · 1 yr 1 mo · Dubai, Dubai, United Arab Emirates

• Built an AI-powered mobile app for visual search of the clothes

Aug 2017 – Feb 2019 · 1 yr 6 mos · Dubai, Dubai, United Arab Emirates · On-site

• Played a key role in the digital transformation of a large retail organization.
• Supported and contributed to the growth of the Analytics team from 3 to over 30 people, providing team mentoring and guidance.
• Championed an experimentation culture within the organization and increased data science awareness through workshops, including initiating a "school of analytics" program.
• Led the delivery of two large data science-centric projects for Carrefour stores, requiring tight cross-functional cooperation between multiple teams and active management of external consultants.
• Designed, developed, and deployed a recommender system for the Carrefour online store from scratch (including data collection, ML models, backend service, and A/B testing). A/B tests demonstrated that the new recommender system increased cross-sell by approximately 130% compared to previous human-curated solutions, showcasing direct business impact.

Aug 2016 – May 2017 · 9 mos · Dubai, United Arab Emirates · On-site

• Contributed to the early growth and development of noon.com, a major Middle East e-commerce platform, in a challenging, fluid multinational environment.
• Built the foundational elements of the recommender system, including designing the user activity data collection pipeline, building a content-based recommendation engine, and designing the A/B testing system.

Jun 2015 – Jul 2016 · 1 yr 1 mo · Moscow, Russia · On-site

• Various data science projects using real "big data" from the second largest social network in Russia (50 million daily active users, ~10 TB daily content).
• Developed a feature extraction pipeline for the feed recommender, increasing user engagement by improving CTR by ~10%.
• Organized a "Social Network Hackathon" competition and follow-up conference, which led to the successful hiring of 3 uniquely skilled ML engineers.
• Assessed the impact of feed CTR on user retention (short/mid/long term) and effectively presented and defended counter-intuitive conclusions to senior management.
• Led the migration of the data science cluster from Hadoop 1.* to 2.0, introducing the YARN resource manager and setting up Spark over YARN.

Dec 2013 – Jun 2015 · 1 yr 6 mos

• As part of the founding team, I’ve been working on creating a service which builds individual's virtual profile based on the digital footprint (social networks, government services, etc) coupled with search engine which could work with incomplete input (for example one can find a person only by providing first name, age range and names of a couple of friends).
• Areas of application: fraud prevention, credit scoring, background check.
• We’ve been able to acquire several mid-size paying customers and attracted a lot of attention from large-sized companies. However, the service was ahead of its time due to obscure regulatory environment in the area of personal data, which restricted our potential.

Apr 2012 – Dec 2013 · 1 yr 8 mos · Moscow, Moscow City, Russia

• R&D on statistical arbitrage trading strategies, backtesting of the strategies
• Developed pricing engine for complex financial derivatives (convertible bonds)
• Designed a Liquidity Index for Russian bond market which was very warmly met by bank clients.