Yassmen Y.

Building AI is easy. Building AI that works at scale is a different problem entirely and that's what I do. With 3 years of experience, I've built systems that go beyond the demo straight to real users, real load, and real edge cases. My clients come to me when they need agent workflows that route and recover intelligently, knowledge systems that retrieve the right answer fast, and conversation flows that hold context across complex multi-turn interactions. I don't just integrate an API and call it a system. I think about concurrency, state persistence, error handling, and what happens when things go wrong at 2am because that's what production actually looks like. My expertise includes: 1. Multi-agent orchestration: LangChain, LangGraph, OpenAI Assistants, custom routing logic 2. RAG systems: hybrid search, pgvector, document ingestion and embedding pipelines 3. Conversational AI: GPT-4, Azure OpenAI, Claude, multi-turn context management 4. Speech AI: real-time speech-to-speech pipelines, speech emotion recognition, audio processing 5. Computer Vision: image processing pipelines, deep learning models, OpenCV 6. OCR & Document AI: text extraction, document parsing, intelligent data capture 7. Backend systems: Python, FastAPI, PostgreSQL, async workflows, Docker 8. Platform integrations: WhatsApp Business API, web chat, messaging platforms 9 Reliability: Langfuse, logging, fallback handling, human-in-the-loop workflows If you need it built properly, let's talk.