Frontend architecture
for products that
need to scale.
I'm Tymoteusz, a frontend engineer with 8 years of experience on large, high-traffic products. Most of my work is architecture, React and design systems.
Where I focus my work
Eight years of frontend work, a good part of it spent on architecture and leading teams. I'd rather know a few things well than a lot of things a little.
Frontend Architecture
Module boundaries, data flow and dependencies set up so the codebase still makes sense two years and ten developers later.
React & TypeScript
Component design, performance patterns and strict typing across big codebases. I care about code that still reads well six months after it ships.
Design Systems
Tokens, component APIs, docs and tooling. The goal is a system teams reach for because it saves them time, not because they're told to.
Performance
Bundle analysis, rendering strategy, Core Web Vitals. I measure first, then fix whatever is slowing users down the most.
Technical Leadership
Engineering standards, code review habits, architectural direction. Mostly it's about making it easier for good engineers to do good work.
Product Engineering
I know enough about product and UX to make sensible tradeoffs. A lot of frontend work happens somewhere between "it works" and "it's right".
AI Product Engineering
Frontends for AI products: streaming UIs, LLM API integration, and a code structure that survives a change of model or provider.
How I engage
Every engagement has a clear scope and a clear result. I only take on a few clients at a time, so each one gets my full attention.
Architecture Review
I go through your frontend codebase and point out structural risks, performance bottlenecks and parts that will be hard to maintain. You get a list of what to fix and in what order.
Embedded Engineering
I join your team and work on the hard frontend problems with you. That means writing code, making architecture calls and raising the bar through reviews.
Design System Build
Tokens, component library and documentation, built around the way your team works instead of copied from a generic template.
Performance Audit
I dig into rendering bottlenecks, bundle size and Core Web Vitals, then rank the fixes by how much they matter to your users. Lighthouse scores are where we start, not the goal.
AI Feature Integration
Getting your AI features into the product UI: streaming responses, tool-call views, structured output and loading states that don't look broken. AI should feel like part of the product, not something bolted on later.
Need something bigger or different? We can work it out.
I build the frontend layer
of AI-powered products.
Streaming interfaces
Token-by-token streaming with sensible loading states, partial rendering and error recovery. The app feels fast even when the model isn't.
LLM API integration
A thin, clean layer over OpenAI, Anthropic and open-source models. Prompts, retries and output parsing live there, away from your components.
AI-native UX patterns
Chat, copilot sidebars, inline suggestions, tool-call visualizations. Designed for AI from the start instead of squeezed into an old UI.
Maintainable as models change
Switching providers or upgrading a model shouldn't break your frontend. I plan the architecture around that from day one.
Works with: OpenAI · Anthropic · Vercel AI SDK · open-source models
Problems I've solved
A few projects I've worked on. Client names are under NDA, but I'm happy to go into detail on a call.
U.S. Enterprise E-commerce Platform
Worked with a U.S. team on rewriting the online store of one of the largest office supply providers in the country. It's used by 300,000 companies and millions of end users, with over 1,000 concurrent visitors at peak. The hard part was customization: every customer can have its own product catalogue, pricing and availability, so the architecture had to handle that from day one.
iOS & Android App (2M+ Downloads)
Worked with a Swedish engineering team on a mobile app with over 2 million downloads. I shipped new features, built shared internal components and took part in product and technical decisions. A lot of my time went into performance and maintainability, including pair programming sessions to spread good patterns across the team.
Technology Audit & Stack Decision
Ran a technology audit for a product team that had to choose a direction for its architecture. I compared frameworks, approaches and tooling against what the team had to work with: expected scale, deadlines and how many people would maintain it. The team based its final stack on those recommendations and avoided months of building on the wrong foundation.
AI-Powered Knowledge Platform
Built the frontend of a B2B knowledge platform with LLM-powered search, streaming chat and structured document generation. I designed the AI layer (streaming responses, tool-call visualizations, fallbacks) so it stayed easy to maintain while the models and providers underneath kept changing.
Technologies I work with
Tools I've used in production over the last 8 years and would happily pick again.
Seniority is about
what you don't
have to explain.
I work best with teams looking for someone who will own the problem, not just close tickets.
Ownership, not task completion
I care about the outcome, not only my own tickets. If I see something wrong nearby, I'll bring it up.
Systems thinking
I design for the fifth feature, not just the first. Decisions made early decide what's possible later.
Honest communication
I don't overpromise. If a deadline is unrealistic or an approach won't work, I'll tell you early and plainly.
Product awareness
I want to know why a decision matters, not just how to implement it. That makes my tradeoffs better.
Clean, reviewable code
Code I write is easy to read, easy to extend, and easy to delete when requirements change.
Deep frontend knowledge
Eight years spent on the hard parts of frontend: architecture, performance and the systems around them.
Tymoteusz Tracz
Frontend Architecture & Engineering
Rzeszów, Poland · Worldwide
For the last eight years I've been building frontends for products where mistakes are expensive: e-commerce platforms, SaaS tools and financial applications. In all of them, performance, reliability and maintainability were hard requirements.
Since 2025 I've also been Head of Frontend Guild at Software Mind, where I set engineering standards, run architecture reviews and help people grow. It's taught me a lot about why some codebases age well and others slowly turn into a burden.
I started CHARTCODE to work directly with companies that want a stronger frontend. I keep the number of projects small, so every client gets my full attention.
Want to raise the bar
on your frontend?
Tell me what you're working on. I'll be honest about whether I can help and how. No strings attached.
I usually reply within 24 hours on business days.