Portrait of Kirsantov

About Me

I'm an AI Product Architect with 10+ years across engineering, product, and design. I explore how LLMs, interfaces, and product systems fit together — and turn those experiments into working products.

Currently CIO at Kyrgyz Single Window (KSW). Based on UTC+6, working with clients globally.

Snapshot

  • 10+ years in IT — development, product, UX/UI
  • AI Product Architect — bridging LLMs, interfaces, and product systems
  • Design systems & components — Storybook, UI kit, interface unification
  • Modernisation and support of complex / legacy systems — refactoring, performance, scaling
  • AI in products — OCR, LLM integrations, and AI-assisted workflows

Experience

  • 2025 – present
    Chief Information Officer — Kyrgyz Single Window (KSW), Bishkek
    • Leading IT direction
    • AI implementation for Customs and Single Window systems
  • 2022 – 2025
    Senior Software Engineer — Social Discovery Ventures
    • Legacy product support and modernisation
    • Participated in redesign: modern framework selection, UIKit
    • Implemented Storybook and a component-based approach
    • Refactoring and performance improvements
    • Participated in launching a new product based on the core platform
  • 2018 – 2022
    Frontend Developer — iBride.com
    • Built a React Native version from scratch for Nevesta.info
    • Maintained the legacy site and Cordova version
    • Gradual migration to a modern stack
  • 2019 – 2020
    Frontend Developer — InvestEngine (London, UK)
  • 2018
    Frontend Developer (React) — Vinza.net (London, UK)
  • 2017 – 2018
    Front-End Developer (React) — Nebeus (London, UK)
    • Crypto service web interface, integrations, UX/UI
  • 2015 – 2018
    Founder — Pinflorist LLC (Riga, Latvia)
  • 2013 – 2018
    Head of Business Development — ZenMall
  • 2012 – 2013
    Art Director — Trade&Chat
  • 2011 – 2012
    Founder — Addqd

Education

  • 2020 – 2024
    University of London — Computer Science

How I work

  1. Understand the real problem before writing code
  2. Find the simplest solution that actually ships
  3. Deploy it, measure it, and learn from real usage
  4. Iterate based on data, not opinions