Senior Frontend / Full Stack Engineer
Oleksa Tyshchenko
10+ years building low-code platforms, back-office systems and AI-powered products.
- dev@oleksatyshchenko.com
- Phone
- +380 73 407 4118
- linkedin.com/
in/ oleksa-t-90a050a8 - GitHub
- github.com/
froii - Website
- oleksatyshchenko.com
- Format
- Remote, in the team's time zone · B2B or EOR · open to relocation
- Languages
- English (B2, professional working proficiency) · Ukrainian
About
I started on the frontend, took on the backend, and now ship LLM features in production. Over ten years, the stack has changed more than once; what carries over is understanding how the whole system fits together. Most of what I ship is infrastructure other teams build on: low-code platforms, back-office systems, document management, e-commerce. I own architecture and performance on large codebases, and run the code review that keeps them maintainable. React, Next.js and TypeScript on the front, Node.js and PostgreSQL behind it.
Key achievements
- Designed a DSL for a low-code reporting platform, so analysts write formulas themselves instead of filing tickets.Wrote the grammar, the editor and the evaluator: a parser built with Chevrotain and Lezer, wired into a CodeMirror editor with completion and inline errors.
- Worked directly with clients in the EU and the US: from deciding what to build, through the trade-offs, to what shipped.Small teams, so I was responsible for the frontend, the backend, the release and more.
- Built premium subscriptions for a PWA that also runs in native apps: a paywall plus store purchases and restores.Purchases go through a native bridge to the store, while premium banners and promo-code and referral rewards live in the web app.
- Took LLM features from prototype to production.Built retrieval over the product's own documents, streaming responses, tool calling and agent-style flows. All of it runs through OpenRouter rather than one vendor.
- Kept large tables and charts responsive on codebases too large for one engineer to keep in their head.Virtualized the big tables, split and lazy-loaded the routes, trimmed the bundle, and cut repeat GraphQL round trips with fragment-level caching.
- Set up the code review process and mentored the developers who joined later.Wrote the specs the code was built from, and the shared rules and prompts the team used with AI coding tools.
Beyond the code
Most of my career has been remote, working with German, American, Israeli and Ukrainian teams. The same sentence carries different weight depending on who says it: direct feedback that is normal in one team reads as rudeness in another. Erin Meyer's "The Culture Map" is good on this. So I keep in mind who I am talking to and adapt how I communicate to the team.
I like difficult problems and games such as Dark Souls. My approach to both is the same: first understand how something works.
Outside work: manga, long walks and tea.
Skills
- FrontendReact · Next.js · TypeScript · GraphQL · Apollo Client · PWA· JavaScript · HTML5 · CSS3 · SCSS · CSS Modules · styled-components · Ant Design · Storybook · React Router
- StateRedux · Zustand · React Query· React Context · Redux-Saga
- BackendNode.js · Express · PostgreSQL · NestJS · TypeORM · REST · API design · WebSockets · Redis· Apollo Server · type-graphql · Axios · Nodemailer
- AI & LLMLLM APIs · RAG · Prompt Engineering · AI Agents· OpenRouter
- ToolingDocker · AWS (Cloud Essentials) · Jenkins · Git · ESLint· GitHub Actions · Webpack · Vite · Prettier · Jest · Vitest · React Testing Library · Enzyme · Figma · Jira · Scrum
- LibrariesCodeMirror 6 · Lezer · Chevrotain · Lexical · Chart.js · react-dnd · react-window · D3.js · Lodash · Dexie · Puppeteer
Experience
- Project
- A social app built around challenges between users: a Next.js PWA with LLM features that also runs inside the iOS and Android apps.
- Result
- Built the client side of the premium subscription and put the product's first LLM features into production.
- Challenges
- RAG without a vector database: for a small set of product documents, keyword search was enough, the same grep-style approach coding agents use.
- A PWA inside the iOS and Android apps: push notifications, local data in IndexedDB through Dexie, and store purchases through a native bridge, built with the team.
- Scope
- The whole frontend of the PWA: challenges, a feed, push notifications, profiles and a rewards wallet.
- An AI assistant panel: streamed answers, model choice, tool calling and undoable AI edits. Wrote the spec before the code.
- Stack
- Secondary
- Project
- nedyx - a low-code platform for business reporting, built by a small international team.
- Result
- Shipped the widget system, undo/redo, archives and PDF export that the platform's reporting runs on, plus the platform's own functional language for formulas.
- Challenges
- Undo/redo for the whole app, written in-house for performance: widget moves, formulas, styles and other changes share one history, a change across several pages undoes as one step, and the history is kept in IndexedDB so it survives a reload.
- A formula editor for analysts rather than developers: references to widgets and cells are highlighted and autocompleted from page and widget names, errors show as you type and after server validation, and clicking a reference selects the widget on the page.
- User-built pages of tables, charts and text on data from SSAS, OData, Google Sheets and external databases: custom drag-and-resize, alignment guides and Puppeteer PDF export work for any set of widgets.
- Scope
- Maintained and extended the product in React, TypeScript, Node.js, GraphQL and Apollo. Built a dynamic widget system with full visualization control, large tables, charts and a text editor, reading data from external databases for many users.
- Stack
- Secondary
- Project
- Three applications for LinkSquares: contract lifecycle and legal project management.
- Result
- Delivered three LinkSquares applications; the dynamic form and table layer became the base the later features were built on.
- Scope
- Built the logic for rendering dynamic and asynchronous forms, tables and components, plus PDF editors, working from Figma and a kanban board. Used React with hooks, GraphQL, Apollo Server, Redux, Redux-Saga, Ant Design, Ramda and several form libraries, alongside my own utilities.
- Stack
- Secondary
Earlier and non-development roles
- Project
- An introductory frontend course.
- Result
- Took a group of 25 beginners from nothing to a working knowledge of HTML, CSS and Git.
- Scope
- Lectures, hands-on sessions and homework review across the full course.
- Stack
- Project
- A site generator: a back office whose settings define how another site is displayed. Alongside it, an online store for leisure events and a TED-style speaker platform in London.
- Result
- Delivered the client side and the back office of a generator that builds and runs other sites.
- Scope
- Email templates and business logic across a large admin panel, with tables and charts for creating dynamic pages, forms, info blocks and galleries.
- Stack
- Secondary
- Project
- The company site and product pages, including Family Tree Maker.
- Result
- Kept the site current through product launches, adding the pages and questionnaires each one needed.
- Scope
- Extended the company's sales site: jQuery on the frontend, PHP on the backend.
- Stack
- Secondary
- Project
- Part-time work through Upwork.
- Result
- Delivered complete sites and mobile versions for clients I found and managed myself.
- Scope
- Everything from individual page components to full sites and mobile-version updates.
- Stack
- Project
- Radioecological and chemical pollution monitoring of water bodies, alongside graduate school and the first years of web work.
- Result
- Published research on water pollution monitoring.
- Scope
- Monitoring and analysis of pollution data, with ArcMap for the spatial work.
- Stack
Earlier roles
- Frontend Developer -Adraba, Kyiv · 01/2018 - 07/2018
- Web Software Developer -Software MacKiev, Kyiv · 09/2016 - 01/2018
- Freelance Frontend Developer · 04/2015 - 10/2016
Education
MSc in Computer Science, AI & Machine Learning
09/2025expected 09/2027
Neoversity IT University
Machine learning, neural networks, statistics and algorithms.
Machine learning · Neural networks · Algorithms & data structures · Python for data · Statistics · Research methods
Postgraduate studies, Environmental Safety
20122016
National Academy of Sciences of Ukraine
Ran in parallel with research work at the Academy: pollution monitoring, data validation and published papers.
Research design · Data quality & validation · Technical writing · ArcMap / GIS
Master's degree, Project Management
20112012
National University "Kyiv Aviation Institute"
Took part in student conferences.
Project management · Planning & estimation · Risk management · Global sourcing · Stakeholder communication
Bachelor's degree, Hydrometeorology
20072011
Taras Shevchenko National University of Kyiv
Physics, statistics and modeling of complex systems. Played chess for the university team.
Data analysis · Statistical modeling · Scientific method · GIS · Large datasets · Self-directed learning
Certifications
Recruiter Q&A and an AI chat: oleksatyshchenko.com/en/questions
Full profile on LinkedIn: linkedin.com/in/oleksa-t-90a050a8
Contact
Tell me what you are working on and how I can help.
Contacts
dev@oleksatyshchenko.com