Vladimir S.

I build production AI systems: LLM applications, RAG platforms and AI agents that keep working after the demo. 10+ years in software engineering, Expert-Vetted on Upwork, and personally involved from the first call to release. Most teams come to me with one of three things: an AI feature for an existing product, an LLM prototype that has to survive real users, or a manual process they want an agent to take over. The difficult work is rarely choosing a model. It is designing the system around it: company data, context engineering, integrations, evaluation, monitoring, latency, cost control, security and predictable behaviour. What I build: ✅ AI agents and workflows: tool use, MCP integrations, memory and context management, human approval gates ✅ RAG and internal knowledge systems: hybrid retrieval, pgvector/Qdrant, reranking, evaluation ✅ LLM applications on OpenAI, Claude and self-hosted models ✅ Real-time voice agents: speech-to-text (ASR) and text-to-speech (TTS) ✅ Computer vision and machine learning for production environments ✅ Production backend: Python, FastAPI, Docker, APIs, AWS/Azure How I keep it reliable: observability, evals, regression testing and audit trails, so you know what the system does and what it costs before your users find out. Recent work: ✅ RAG troubleshooting system for a UK manufacturer, funded by Innovate UK ✅ Production AI and computer vision for electronics manufacturing ✅ AI investment-research platform ✅ LLM compliance automation for AdTech ✅ AI-assisted music-therapy product used in dementia and autism care Still deciding where AI fits? I also work as a fractional AI lead: mapping workflows, prioritising use cases, technical due diligence, build-vs-buy decisions, and testing the direction before a team commits serious budget. I'm the founder of Go Wombat, an AI engineering company, and I run AI Talks Valencia, the largest local AI community in Valencia. Send me a short note on what you're building and where it's stuck. I'll reply with how I'd approach it.