Aiman Shahin

Consultant @Machine Learning Reply GmbH

Berlin, DE

Designed and implemented an intelligent Retrieval-Augmented Generation (RAG) chatbot system powered by GPT-4 with OpenAI embeddings and FAISS, introducing a novel architecture where each document is mapped to its own vector space; through an initial semantic reasoning layer, the system dynamically determines whether to focus on a single document or multiple documents, enabling adaptive vector-space switching and delivering smarter, context-aware responses in real time via a FastAPI backend with Server-Sent Events and AI routing. This helped in filtering between documents with very similar content.• Led a machine vision initiative, designing, training, and deploying AI inspection models with Vertex AI, Google Cloud Console, Label Studio, and Google Colab to automate quality control and improve operational efficiency.• Developed and deployed a live-updating Power BI dashboard integrated with Google BigQuery, using TMDL, DAX, and custom scripts to provide real-time business intelligence and data-driven decision-making for stakeholders.• Created a Python-based compliance framework aligned with the EU AI Act, embedded into an AI agent to automatically validate outputs and monitor code repositories for regulatory adherence, reducing compliance risks.• Built and maintained Generative AI pipelines with ComfyUI to produce high-quality, brand-consistent marketing images, accelerating creative workflows and supporting corporate marketing campaigns.