Viktor Turskyi

Українською

Technology executive and software architect · Kyiv, Ukraine

contact@viktorturskyi.com

I co-founded the software company WebbyLab in 2011 and grew it from three people to 150, first as CTO and then as CEO. Over ten years we delivered more than 60 complex projects, five of them for Fortune 500 companies. From 2021 to 2025 I was a team lead at Google, working on Google Cloud Run.

I now advise companies on architecture, engineering leadership and applied AI. I am also a PhD candidate at Igor Sikorsky Kyiv Polytechnic Institute, where I research how language models can extract sensitive information from large volumes of text. I started out in information security in 2003 and have been writing software ever since.

Explaining things is a big part of my work. I have given more than 50 conference talks, run a Ukrainian YouTube channel about programming fundamentals that won the community vote at the first DOU Award, and teach at American University Kyiv.

Consulting

I take on a small number of engagements, remotely or in Kyiv, in English or Ukrainian. NDAs are fine.

  • Fractional CTO. Technical leadership for companies that don’t have a CTO yet, have just hired one, or have outgrown their processes. Usually one or two days a week.

    • Technology strategy and a roadmap tied to business goals
    • Team structure, hiring process and a skills baseline for engineers
    • Delivery process, key architecture decisions, technical conversations with investors and clients

  • Architecture review and technical due diligence. An independent look at a system before scaling, a rewrite or an investment. Fixed scope, one to three weeks, ending with a written report and a debrief.

    • Architecture, code quality, infrastructure and delivery process
    • Scalability, performance and security risks, and dependence on key people
    • A prioritised plan with effort estimates

  • AI and LLM integration. For teams that want language models to do real work in a product or in the engineering process, beyond demos.

    • Choosing the use cases where AI pays off
    • Designing LLM features: structured extraction, retrieval, knowledge graphs, evaluation
    • Handling sensitive data, prompt injection and access control; AI-assisted development with clear rules

To discuss a project, write to contact@viktorturskyi.com. The first call is free.

Expertise

  • Leadership. C-level management: company and product strategy, building and optimising development processes, hiring and mentoring; grew a company from 3 to 150 people.
  • Architecture. Designing complex software systems, building high-performance engines and developing secure systems; the architecture of 50+ projects lasting from two months to three years.
  • AI. Fifteen years of interest in AI: from Stanford’s 2011 course, a chatbot’s NLP core and a question-generation engine before the LLM era to PhD research in LLMs and knowledge graphs.
  • Security. From PKI and security audits in 2003 to an end-to-end encrypted email system, application security training and research in cyber threat intelligence.
  • Technologies. JavaScript and TypeScript, Node.js, React and React Native, Perl, Go, Flutter and Dart, Linux, Google Cloud.

Experience

  1. 2026

    Visiting Lecturer, American University Kyiv

    Course “Declarative Frontend Programming with AI” at the EPAM School of Digital Technologies.

  2. 2021 – 2025

    Team Lead, Senior Software Engineer, Google

    Google Cloud Run, the serverless platform of Google Cloud.

    • Led an engineering pod and aligned the technical vision across frontend, backend and AI teams
    • Planned development stages with the product team, developed the product architecture and approved every architecture plan for the team
    • The product launch was featured in the keynotes and in five talks at Google Cloud Next 2024
  3. 2021 – now

    Non-Executive Director, WebbyLab

  4. 2011 – 2021

    Co-founder, CTO, then CEO and Chief Architect, WebbyLab

    • Grew the company from 3 to 150 people; set the company strategy and introduced engineering practices that made it more competitive
    • Delivered more than 60 projects (React, React Native, Node.js, NLP, Hadoop, Perl), including work for five Fortune 500 companies
    • Designed the architecture of 20 long-term (15–35 months) and more than 30 shorter projects; wrote code for the most critical subsystems
    • Scaled delivery to large corporate clients with teams of up to 15 engineers; started the IoT, smart-home and blockchain practices
    • Prepared technical and financial proposals, taught software architecture to clients’ teams and mentored more than 50 engineers
  5. 2016 – 2021

    CTO, ITSQUIZ

    Online testing platform.

    • Designed the architecture of a natural language processing engine that generates quizzes
    • Hired and managed the technical team; wrote code for the critical subsystems
  6. 2015 – 2016

    JavaScript Trainer, GoIT

    Taught the Advanced Frontend course.

  7. 2011

    Software Engineer, PortaOne

    Software for VoIP providers and telecom operators.

    • Developed and optimised the Porta Billing system (Perl, RADIUS, SIP, H.323, RTMP)
    • Made the Porta-Monitor interface, which shows data from 1,500 servers, 20 times faster
    • Analysed and planned disaster recovery for Porta Switch
  8. 2007 – 2011

    Webmail Developer, Web Team Lead, Lead System-Level Developer, Mail.ua

    A free national email service in Ukraine.

    • Built a fully AJAX webmail in JavaScript and Perl, with a virtual-scrolling grid for more than a million rows
    • Led the webmail team of five, then the system team of three; moved every component to a single distributed storage layer covered by integration tests
    • Improved the web framework (request dispatcher, sessions, FastCGI workers), built deployment pipelines and a release process that made rollouts predictable; trained ten interns
  9. 2003 – 2007

    Chief System Administrator (volunteer), Taras Shevchenko University dormitory

    Built a network of more than 200 nodes from scratch, from cabling to Linux servers; coordinated four administrators and automated switch management with Perl.

  10. 2003 – 2004

    Intern, Information Security Department, BMS Consulting

    Public key infrastructure and security audits; certified as a Microsoft Certified Professional.

Selected client work

  • Spreadsheet engine. A calculation engine written entirely in JavaScript that runs Microsoft Excel files in the browser or on the server, optimised for up to a million cells. Used for heavy computations by some of the world’s largest pharmaceutical companies.
  • Presentation platform. A Keynote and PowerPoint alternative with web, mobile and desktop apps, used by several of the five largest pharmaceutical companies.
  • Insurance chatbot. An AI-driven assistant for a San Francisco startup; I designed its natural language processing core.
  • Social media monitoring. A Hadoop cluster on AWS EMR, collectors for social network APIs, a task queue and a backend for analysing the processed data.
  • Track-day marketplace. A marketplace for a performance car brand where racers and enthusiasts rent cars for races and track days around the world.
  • Accounting system. An online accounting product with a REST API and a rich JavaScript interface: more than 80,000 lines and 800 classes in Perl and JavaScript.
  • Smart home platform. A modular automation platform that brings devices from different manufacturers and protocols into one system.
  • Encrypted email. A secure email system with end-to-end encryption, similar to Proton Mail.

Products

  • 2026

    CamCtrlcamctrl.app

    Remote control for Blackmagic cameras from a phone, tablet or browser.

    Focus, exposure, audio, media, monitoring and colour over the local network. Flutter, Blackmagic REST API, WebSocket. Web app available; iOS and Android coming.

  • 2026

    NoTeleportnoteleport.github.io

    Keeps Android navigation working in Ukraine when GPS is spoofed.

    Detects impossible fixes and switches GPS off for every app until spoofing ends. Kotlin and Flutter; no internet permission, no data collection.

  • 2026

    VidCtrlcamctrl.app/vidctrl

    Desktop video converter on FFmpeg: pick a goal, not codec settings.

    Archive masters in AV1 or HEVC with metadata kept, DaVinci Resolve preparation on Linux, phone-friendly copies, metadata removal. Flutter; Linux, macOS, Windows.

  • 2026

    Type Racingssh type.tvvracing.com

    A code typing race that runs entirely in the terminal over SSH.

    Three weekly tracks, ghost replays of real runs, and the SSH key as the account. Go, Wish, Bubble Tea, MariaDB.

  • 2021

    my-talks.netmy-talks.net

    A platform where speakers collect their talks, slides and videos.

    Speaker profiles, events and calls for papers. Node.js, Next.js and MySQL on Google Cloud, built on my own open-source libraries.

Open source

  • 2012

    LIVRlivr-spec.org

    Language-independent validation rules: describe validation once as data and run it in any language.

    Specification (291 stars) and JavaScript library (183 stars, about 9,000 npm downloads a week); implementations in about a dozen languages.

  • 2018

    Mole RPCgithub.com/koorchik/node-mole-rpc

    Transport-agnostic JSON-RPC 2.0 client and server for Node.js and browsers.

  • 2018

    Chistagithub.com/koorchik/node-chista

    A toolkit for building Express services with clean architecture.

  • 2019

    ConfMegithub.com/koorchik/node-confme

    A configuration library for complex configs that follows the twelve-factor app rules.

  • 2026

    dom-harness, mui-harnessgithub.com/koorchik/dom-harness

    Framework-agnostic test harnesses for UI components, with a Material UI layer.

  • 2010 – 2018

    CPANmetacpan.org/author/KOORCHIK

    18 Perl distributions, including Validator::LIVR and Mojolicious plugins for CSRF protection, session-to-IP binding and logging.

Research

My PhD research is about turning unstructured, sensitive text, such as cyber-incident reports, into knowledge graphs that analysts can verify.

Dissertation: Methodology for Extracting Sensitive Information from Large Text Arrays Based on Artificial Intelligence.
Focus: Formal mathematical models for building knowledge graphs from unstructured text; large language models for cybersecurity; analysis and automated processing of cyber threat intelligence.

  1. 2026

    Streaming Construction of a SKOS-Aligned Concept Registry from Cybersecurity Incident Reports

    Theoretical and Applied Cybersecurity, 8(2).

    A streaming method that keeps a normalised, hierarchical concept registry current as reports arrive. On 204 CERT-UA reports an open-weight model stack matched a cloud frontier stack.

  2. 2025

    A Formal Model for Constructing Sensitive Data Graphs from Cyber Reports using Large Language Models

    Theoretical and Applied Cybersecurity, 7(2), 98–107.

    A four-stage formal model that combines LLM extraction with deterministic rules and network analysis to build canonical knowledge graphs from 204 CERT-UA incident reports.

Profiles: ORCID · Semantic Scholar · OpenAlex

Speaking and teaching

  • Talks. More than 50 since 2011 in Ukraine, Norway, the Netherlands and Latvia, including JSNation (Amsterdam), Mobile Era (Oslo), PerlCon (Riga), JavaScript fwdays, JS Fest, OdessaJS, KharkivJS and DOU Day. Topics: software architecture and project risk, Node.js and frontend engineering, application security, engineering teams and careers. Videos and slides
  • YouTube. “Віктор Турський про програмування”, a Ukrainian channel on programming fundamentals: how the internet, databases and operating systems work, algorithms, security, and live coding with AI. 21.7K subscribers and 435K views. Channel
  • Telegram. “Жабаскрипт”, short posts on architecture, best practices, performance and security, 4.7K subscribers. Channel
  • Teaching. Visiting Lecturer at American University Kyiv; lectures on software architecture at Kyiv School of Economics and AUK; online workshops on Google Cloud; the Advanced Frontend course at GoIT. Mentored more than 50 engineers.

Recognition

Interviews and press

Education

  1. 2024 – now

    PhD studies in AI and cybersecurity

    Igor Sikorsky Kyiv Polytechnic Institute, Institute of Physics and Technology, Department of Information Security

  2. 2020

    Application Security

    In-person course, Techmaker

  3. 2019

    Embedded Basic

    In-person course, Techmaker

  4. 2011

    Introduction to Artificial Intelligence

    Online course with Stanford Engineering, final score 95.7%

  5. 2005 – 2007

    Master’s degree in International Economics

    Taras Shevchenko National University of Kyiv, Faculty of Economics

  6. 2001 – 2005

    Bachelor’s degree in International Economics

    Taras Shevchenko National University of Kyiv; exchange student at Fachhochschule Schmalkalden, Germany, in 2003

  7. Microsoft Certified Professional

    Windows 2000 Server, Active Directory and network administration

  8. Languages

    Ukrainian (native), English (advanced, C1).

Contact