Saad S.

I build production RAG and AI-agent systems that turn documents, databases and business APIs into accurate, source-cited workflows. Using Python, FastAPI, LangGraph and vector search, I help SaaS and enterprise teams move from fragile AI demos to reliable products. My focus is not simply connecting an LLM to a chatbot interface. I design the complete backend and retrieval workflow: document ingestion, chunking, metadata, embeddings, semantic or hybrid search, reranking, grounding, citations, agent orchestration, API integrations and production deployment. I can help you build or improve: • Retrieval-Augmented Generation systems • AI agents and tool-calling workflows • Document AI and PDF-processing pipelines • Enterprise knowledge assistants • Semantic and hybrid search applications • Source-cited research and compliance assistants • Internal support and knowledge-base chatbots • Human-in-the-loop review workflows • Python and FastAPI AI backends • Multi-source data and API integrations Typical problems I solve include: • Relevant information exists in the documents, but the RAG system fails to retrieve it • Answers are not grounded in the approved knowledge base • Citations point to the wrong source or document section • Chunking, metadata or embeddings reduce retrieval quality • An AI-agent prototype becomes unreliable when tools, branching and retries are added • A no-code workflow needs to be rebuilt as a maintainable Python service • An AI application needs secure APIs, logging, error handling, deployment and documentation My technical stack includes Python, FastAPI, LangGraph, LangChain, LlamaIndex, OpenAI, Claude, Qdrant, Pinecone, Weaviate, ChromaDB, PostgreSQL, Elasticsearch, Docker, AWS and Azure. I have more than six years of software-development experience across AI engineering, backend systems, data pipelines and SaaS applications. I am also a Microsoft Certified Azure AI Engineer Associate and an AWS Certified Machine Learning – Specialty. On Upwork, I have completed backend AI automation and product-data engineering engagements. My AI automation contract received five-star feedback recognising my communication, technical skills, collaboration and problem-solving. For a new engagement, I can support architecture, implementation, retrieval evaluation, debugging, deployment and ongoing optimisation. I communicate technical trade-offs clearly, document the system and build for maintainability rather than producing a one-off demo. Send me your current architecture, sample documents or the main failure cases. I will identify the highest-risk part of the system and recommend a practical first step.