Mykola Dronov Contact

Mykola Dronov

AI-First Full-Stack Developer

Full-stack developer building production web applications and AI-powered products.

Mykola Dronov

Expertise

About

I’m a full-stack developer focused on building production web applications and AI-powered products. I specialize in TypeScript, React, Next.js, AWS and Node.js, with experience delivering complex systems across e-commerce, payments, data-heavy interfaces, platform integrations, and performance-critical applications.

In recent years, I’ve expanded my focus into AI engineering, building LLM-powered systems that go beyond simple chat interfaces. My experience includes AI agents, multi-agent workflows, RAG, tool-using agents, structured LLM pipelines, API integrations, and automation that connects models to real business data and actions. I’m particularly interested in making AI reliable in production through evaluation, observability, guardrails, and well-designed workflows.

I enjoy taking ambiguous problems from idea to production: shaping architecture, building polished user experiences, designing APIs and integrations, working with cloud-native infrastructure, and continuously improving performance, reliability, and developer experience.

Current role
Senior Software Engineer at Breitling
Based in
Warsaw, Poland
In production since
Mar 2020
Focus
AI agents and LLM pipelines
Payments and checkout
Data-heavy interfaces

Experience

  1. May 2026 – Present

    Senior Software Engineer

    Breitling

  2. Apr 2023 – May 2026

    Software Engineer

    BigCommerce, Warsaw

    • Part of the Payments group integrating and maintaining payment methods across multiple surfaces.
    • Integrated and supported PayPal, Google Pay, Apple Pay and Stripe, ensuring reliable checkout coverage across new and existing payment flows.
    • Implemented and supported the Stripe Express Checkout element within the platform, including UX states and edge-case handling.
    Show 2 more
    • Contributed to payment flow reliability and test coverage with Jest and React Testing Library, improving confidence in critical transactions.
    • Collaborated across sub-teams to scope integrations and decompose work, coordinating with backend and platform stakeholders; touched PHP for client-side needs.
  3. Feb 2022 – Apr 2023

    Software Engineer

    EPAM Systems, Ukraine

    • Worked on a pharmacy platform as part of a development team, owning complex data-heavy UI and performance optimization for tabular workflows.
    • Contributed to a custom virtualized table that renders thousands of rows and hundreds of columns without jank, using react-virtualized, sticky/large cells and memoization.
    • Designed and published a company-wide UI npm library from scratch, so multiple projects reuse the same primitives and design.
    Show 3 more
    • Worked with Redux and Redux-Saga flows for data fetching, state orchestration and error handling across table interactions.
    • Added tests with Jest and React Testing Library, covering critical components and data flows.
    • Worked closely with QA and backend to profile and resolve bottlenecks in rendering and data hydration for large datasets.
  4. Mar 2020 – Feb 2022

    Front-end Developer

    SeaRates.com, Odesa, Ukraine

    • Worked on Logistics Explorer, a cargo tracking web app, delivering responsive, data-driven UI modules as part of a development team.
    • Built responsive React UI with TypeScript and Redux, with GraphQL data fetching and error handling for shipment and route data.
    • Implemented Leaflet maps with markers and route overlays, plus debounced search with autocomplete to find locations and render routes from A to B.
    Show 3 more
    • Built a fast search bar using request debouncing and incremental rendering.
    • Set up Storybook component docs and covered key components with Jest and Enzyme tests.
    • Automated build and test pipelines with GitHub Actions for the team's releases.

Projects

Products I built and run outside my day job.

Itsbundle

Marketing packs for real-estate photographers, generated from listing photos.

  • Next.js
  • Clerk
  • Stripe
  • Prisma
  • Supabase
  • Fal.ai
  • Flux 2
  • Nano Banana 2
  • Kling
  • MiniMax
  • FFmpeg
  • Supabase Storage
How it works
  1. Listing photos

    A photographer uploads a shoot, tags each room and picks a style.

    Next.js, Clerk

  2. Checkout

    After payment, the job is queued in Postgres for a worker.

    Stripe, Prisma, Supabase

  3. Generate

    The worker stages rooms, makes dusk exteriors, animates 5-second clips and records a voiceover.

    Fal.ai, Flux 2, Nano Banana 2, Kling, MiniMax

  4. Assemble

    Clips and voiceover are stitched into a tour video.

    FFmpeg

  5. Marketing pack

    Stills, clips, the tour and a labeled ZIP, behind a share link.

    Supabase Storage

Next.js 15, React 19, TypeScript and Tailwind CSS. Runs on Docker and k3s, deployed by GitHub Actions. Claude drafts the blog articles.

AI Для Своїх

A Ukrainian Telegram channel about AI tools, run entirely by a bot.

  • GitHub Actions
  • feedparser
  • BeautifulSoup
  • Claude Haiku 4.5
  • Telegram Bot API
  • Git
How it works
  1. AI news

    A GitHub Actions cron job reads RSS feeds three times a day.

    GitHub Actions, feedparser

  2. Filter

    Strips the HTML and keeps the items that match a keyword list.

    BeautifulSoup

  3. Write

    Rewrites each item as a post in Ukrainian, translating English sources.

    Claude Haiku 4.5

  4. Publish

    Two posts go out to the channel on every run.

    Telegram Bot API

  5. Remember

    Published links are committed back to the repo, so a story never posts twice.

    Git

A single Python script. There is no server or database.

Presetti

A desktop app that searches synth presets by how they sound instead of by name.

  • C++17
  • JUCE 8
  • FLAC
  • Python
  • librosa
  • NumPy
How it works
  1. Your synths

    Scans the computer for supported VST3 and VST2 synths.

    C++17, JUCE 8

  2. Render

    Loads each synth, applies its presets and records a short clip of each.

    JUCE 8, FLAC

  3. Index

    A Python sidecar extracts audio features and builds a similarity index.

    Python, librosa, NumPy

  4. Search

    Record into the mic, drop in a sample, or pick a preset to find more like it.

  5. Matches

    Similar presets from the whole library, offline and without an account.

The site at presetti.app is Next.js 16 with a WebGL shader background and an early-access list on Supabase and Resend.

Skills

AI engineering
  • AI agents
  • Multi-agent systems
  • Retrieval-augmented generation (RAG)
  • Model Context Protocol (MCP)
  • LLM pipelines
  • Vector databases
  • Amazon Bedrock
  • Claude API
  • Evals
  • Claude Code
  • OpenAI Codex
Frontend
  • React
  • Next.js
  • TypeScript
  • JavaScript
  • Redux
  • Redux Toolkit
  • React Router
  • Custom hooks
  • GraphQL
  • HTML5
  • CSS
  • SCSS
  • Tailwind CSS
  • styled-components
  • Storybook
  • Webpack
  • AJAX
Backend and cloud
  • Node.js
  • PHP
  • SQL
  • AWS
  • AWS Lambda
  • Amazon API Gateway
  • Amazon S3
  • Amazon CloudWatch
  • Supabase
  • Prisma
  • Docker
  • Kubernetes
  • Object-oriented programming
Payments
  • Stripe
  • PayPal
  • PayPal Checkout SDK
  • Apple Pay
  • Google Pay
Testing and delivery
  • Jest
  • React Testing Library
  • Git
  • GitHub Actions
  • CI/CD
  • npm libraries
  • Performance profiling
Search and compliance
  • SEO
  • Generative engine optimization (GEO)
  • GDPR

Contact

Email is the fastest way to reach me.