Glib S.

Full-stack AI engineer for production SaaS. I build tool-using AI agents, secure RAG, document automation and data pipelines with React, Node.js and Python, and take them from idea or fragile prototype to a product people rely on. $170K+ earned on Upwork | 2,500+ hours | 5.0 from 32 reviews | 8+ years of engineering "One of the best AI Engineers we've ever worked with." (Upwork client, 2026) WHAT I BUILD - AI agents that use real tools, APIs and internal data, and return cited answers instead of chat demos. - Secure RAG and knowledge systems: permissions enforced in the database, citations, evaluation and fallback behavior. - Document AI that turns PDF, Word and Excel files into structured, reviewable data. - Web scraping and data extraction at scale, with LLM parsing where rule-based parsers break down. - Complete AI SaaS products: frontend, backend, authentication, billing, admin tools and deployment. - Rescue work for existing or AI-generated codebases that fall apart past the demo. RESULTS FROM RECENT WORK - Partner-discovery SaaS: a Claude tool-use agent that researches and qualifies 10,000+ companies and returns cited briefs. - Public-records data platform: extraction agents for 28 court jurisdictions across 12 US states; LLM parsing recovered sources the team had written off. - Supplier document AI: retailer form completion cut from about 8 hours to under 5 minutes (client-published figure). - Listing automation for an e-commerce seller: saved the client 80–90 hours of manual work every week. - Property data platform: 2.36M properties linked from 20 government data feeds, with ML valuations at 5.4% median error. - Mobile app rebuild: after relaunch the product passed 5,000 active users and raised $100K+. HOW I WORK I start from the business outcome and the real constraints: your data, user roles, security, latency, cost, and who will operate the feature after launch. You get working increments, direct communication, clear tradeoffs and maintainable code. Based in Ukraine (UTC+3), overlapping with EU working hours and US mornings. CORE STACK TypeScript, React, Next.js, Node.js, NestJS, Python, FastAPI, PostgreSQL (pgvector), MongoDB, AWS, Docker. AI: OpenAI, Claude, Gemini, LangChain, embeddings, vector search, tool calling, structured outputs, MCP, and computer vision when the product needs it. Background: R&D engineer at SoftServe, co-founder and CTO of a product studio, PhD in Artificial Intelligence Systems in progress. Send a short note with what you're building, what exists today, and the outcome you need. For a fixed-scope start, see my project catalog: AI app audit, AI feature architecture sprint, or secure RAG sprint.