About me
Hi, my name is Eugene and I'm a software engineer. I like math, science, algorithms, good UX and other boring things. I started working as a developer in 2011 and received my system engineer diploma from BSUIR in 2013. Most of my experience has been as a fullstack developer (Node/Java + Frontend). I'm especially into functional programming, security engineering, k8s, cloud platforms and graphics work with Three.js/WebGL. Currently, I'm working as a fullstack developer, mainly with Svelte, React, TypeScript, k8s, and AWS/GCP. I enjoy taking messy product and platform problems, building the first useful version, then hardening it into production software that teams can operate and improve.
You can find my detailed CV in pdf here
Details
Focused on
14+ years building production web systems across frontend, backend, infrastructure and integrations.
Hands-on with product delivery: scoping, trade-offs, technical communication and production rollout.
Strong current focus on security, auth/passkeys, k8s, AI agents, Three.js/WebGL, tool execution and observability.
Experience
Recent delivery
Yubico x OpenAI partner launch
Supported Yubico-side delivery for the public OpenAI Advanced Account Security / YubiKey bundle in spring 2026. (public launch)
Result: Shipped partner API integration, a custom product page, custom checkout flow and customer authorization handoff across Yubico ecommerce and account systems.
Customer Accounts / FIDO2 authorization
Work directly with Yubico auth and security teams on an organization-wide authorization flow for future products.
Result: Help scope service details and integrate the flow across product surfaces, ecommerce, account systems and multiple internal teams.
Commerce launches and operational readiness
Delivered customer-facing third-party commerce work across Yubico and Daniel Wellington.
Result: Supported local payment methods, regional store launches, campaign delivery and Black Friday readiness under production launch constraints.
Agentic product systems
Kept a long-running Telegram bot evolving from local chat utilities into a production-like AI system.
Result: Built async workers, reply gating, model fallback, tool execution, memory, metrics and feedback loops around real group-chat behavior.
Careers
[





Education
[
Abilities
Skills
Languages
{Projects
Featured project
A long-running Telegram bot that grew from simple chat commands into an async AWS Lambda system with a reply gate, agent loop, tool execution, memory, metrics and live stats.
Architecture notes on model behavior, tool boundaries, feedback loops and operational patterns.