AI Engineer & Fullstack Developer — LLM orchestration, RAG & AI agents on Vue, Nuxt, TypeScript and Node
I build AI systems that reach production: LLM orchestration and multi-step AI agents, retrieval-augmented generation (RAG) over real company data, and low-code automation pipelines in n8n that replaced roughly 40% of manual internal operations.
The other half of the job is shipping the product around it — 5+ years of commercial fullstack work in JavaScript and TypeScript: Vue 3, Nuxt 3 and React on the front, Node.js, Laravel and Python services, REST and GraphQL APIs, Supabase, Firebase and Strapi behind them. SaaS platforms, CRM systems, browser extensions and cross-platform apps.
Six things I get hired for, as an AI engineer and as a fullstack developer — usually both at once, because an AI feature is only as good as the product wrapped around it.
Multi-step AI agents that call tools, route between models and hand work off to each other. Orchestration layers with retries, guardrails, cost control and evaluation loops — so an LLM feature behaves predictably in production instead of only in a demo.
Retrieval-augmented generation over real company data: chunking and embedding strategy, vector search, reranking and grounded answers with sources. I shipped a RAG-based AI support system that answers and routes requests automatically.
End-to-end features in JavaScript and TypeScript — Vue 3, Nuxt 3 and React on the front, Node.js, Laravel and Python services behind them, REST and GraphQL APIs, Supabase, Firebase, Strapi and PostgreSQL for data. SaaS platforms, CRM systems and dashboards.
Business workflows automated in n8n and connected to LLMs, webhooks, CRMs and internal APIs. Low-code where it ships faster, real code where it has to scale — this cut manual internal operations by about 40% at my current role.
Bringing Cursor, Claude Code and MCP-based tooling into a team's day-to-day: code generation, automated refactors, review assistance and repo-aware agents, with the conventions and guardrails that keep the output reviewable.
Nuxt and Next.js rendering strategy, bundle and Lighthouse work, structured data and technical SEO. Past results include a Nuxt 2 to Nuxt 3 migration worth +25% performance and a 50% cut in initial load time on a legacy CRM.
The stack I use day to day as an AI engineer and fullstack developer.
From AI orchestration and LLM APIs down to the frontend, backend and delivery tooling I ship with.
Languages, education, methodologies, and the tools I use to ship work day to day.
Projects I've built or contributed to. Click to learn more.
Influencer marketing platform where 800+ businesses run campaigns. Creator stats come straight from the social networks' APIs — no screenshots, no bought followers. Ready-made packages with prices, one dashboard and one invoice instead of dozens of chats.
Free VPN trusted by 15M+ users worldwide with 1260+ servers across 60+ countries. I worked across all three surfaces — the marketing site, the Chrome / Firefox browser extension, and the user account dashboard.
Three engineering roles across SaaS, VPN, and CRM platforms — from Vue and Nuxt frontends to AI orchestration and automation in production.
Software Engineer — AI & Fullstack
Frontend Developer
Frontend Developer
I'm an AI Engineer and Fullstack Developer with 5+ years of commercial experience. Most of my work now sits at the seam between the two: LLM orchestration, RAG pipelines and AI agents on one side, and the SaaS product, APIs and interface that have to carry them on the other.
On the AI side I build multi-step agents with tool calling, retrieval-augmented generation over real company data, and automation pipelines in n8n wired into CRMs, webhooks and internal APIs. At Rest&Vest LLC that stack replaced roughly 40% of manual internal operations and now answers and routes support requests through a RAG-based assistant. I also run AI-assisted development in the team itself — Cursor, Claude Code and MCP tooling for generation, refactoring and review.
On the fullstack side my core is JavaScript and TypeScript: Vue 3, Nuxt 3 and React across the frontend, Node.js, Laravel and Python services behind them, REST and GraphQL APIs, and Supabase, Firebase, Strapi or PostgreSQL for data. I've shipped SaaS platforms, CRM systems, browser extensions and cross-platform builds from a single codebase.
Across three engineering roles I migrated a product from Nuxt 2 to Nuxt 3 (+25% performance), localised a platform for Arab and Asian markets (+30% reach), halved initial load time on a legacy CRM, and built a shared UI component library adopted across multiple internal projects. I also mentor — code reviews, pairing, and helping engineers pick up the AI tooling without losing the engineering discipline.
I value clean architecture, maintainable code and product-oriented thinking. Outside of work I run and cycle to stay sharp, and I build small browser games in the open — they're on this site under For Fun. For me, technology is both craft and curiosity.
The questions recruiters and founders ask me first.
In my case: designing the orchestration layer around the model. That means choosing where an LLM belongs and where plain code is better, wiring tool-calling agents, building RAG retrieval over company data, adding guardrails, evaluations, retries and cost limits, and then integrating all of it into the product's existing backend and UI.
Fullstack, with the deepest roots in frontend. Five-plus years of commercial Vue, Nuxt, React and TypeScript work, plus the services around them — Node.js and Express, Laravel and PHP, Python, REST and GraphQL APIs, PostgreSQL, Supabase, Firebase and Strapi. I also ship browser extensions and cross-platform builds with Capacitor, Electron and Flutter.
OpenAI and Anthropic Claude APIs for generation and tool calling, RAG pipelines with embeddings and vector search, n8n for orchestration and automation, and MCP-based tooling with Cursor and Claude Code for AI-assisted development. The glue is usually TypeScript and Node.js.
Yes — n8n is a core part of how I deliver automation, connected to LLMs, webhooks, CRMs and internal APIs. I treat low-code as a delivery speed decision: it carries the workflow until it needs to scale, then the critical path moves into real code without rebuilding everything.
That is a recurring part of my work. I have migrated a production app from Nuxt 2 to Nuxt 3 with a 25% performance gain, moved state from Vuex to Pinia, replaced Webpack and Gulp with Vite for a 50% faster build, and cut initial load time in half on a legacy Vuetify CRM through lazy loading, route-level code splitting and asset optimisation.
Yes — open to fullstack and AI engineering roles, remote across Europe or on-site in Poland, as well as focused AI automation and orchestration projects. The contact form on this page reaches me directly.