Building reliable AI agents, RAG systems, and document automation for real business operations — while preparing the launch of FLOWENTIC
Four years of enterprise automation followed by hands-on agentic AI engineering — LangGraph pipelines, multi-model LLM work (OpenAI, Claude, DeepSeek, Groq), and Human-in-the-Loop systems designed for real operations.
Available for selected contract roles and B2B engagements
// next_chapter
I am currently preparing FLOWENTIC, a B2B AI automation and software company focused on reliable AI agents, RAG, document automation, and enterprise integrations.
The company is being prepared for launch. Until the operational setup is complete, this portfolio remains the primary place to explore my technical work and discuss selected contract engagements.
AI agents for business operations
Auditable multi-agent workflows with Human-in-the-Loop control points.
RAG and enterprise knowledge systems
Hybrid retrieval, document ingestion, and knowledge pipelines for operational teams.
Document and workflow automation
Connecting AI to approvals, tools, and real business process constraints.
// selected_projects
Four technical systems and deployed demonstrations covering agentic orchestration, voice AI, RAG, and operational automation — personal/open-source work, not client production deployments.
LangGraph × Qdrant × Langfuse
My working lab for agentic patterns — LangGraph orchestration, hybrid retrieval on Qdrant, and full Langfuse observability. Four years of enterprise RPA at ATOS shape how I design these systems: for operational constraints and failure modes, not only the happy path.
Chapter V
???
Coming next…// always building
Twilio × LangGraph × ITSM
Pilot-case outbound voice AI: when an ITSM ticket opens, the agent can call the customer, verify identity via a 6-digit code, collect a work note and ETA, then push results back through n8n into ITSM. Designed as a production-grade prototype, not presented as a live client deployment.
FastAPI × Qdrant × Redis × Langfuse
A production-grade, open-source RAG reference implementation — architecture built to production standards, not a claim of client production deployment. Hybrid dense+sparse search with RRF fusion, cross-encoder reranking, async ingest with a DLQ, streaming SSE answers, namespace-scoped RBAC, Prometheus metrics, and per-stage Langfuse tracing across 13 file formats, including audio via Whisper.
// 2026 rag stack
LangGraph × Groq × FastAPI × ChromaDB × SSE
A live deployed demo on Hugging Face Spaces (not a client production system). A LangGraph orchestrator routes each visitor request to the right specialist: RAG for questions about my background, Email to send the CV, Calendar to book a Google Meet, Job Match to score JD fit, and Telegram for instant notifications. Every routing decision streams live to the UI via SSE.
// agents_active
Open Source
All public projects — github.com/georgelush
YouTube
Project demos & walkthroughs
Cybersecurity labs, RPA automation, Unreal Engine 5 prototypes, and C++ systems — the technical foundations behind the agentic systems I build today.