I don’t just want to use an LLM. I want to understand how to build a reliable end-to-end RAG system.

In this video, I present the current state of my RAG project, running locally with Docker.

The MVP already includes:

— an API built with FastAPI; — vector indexing and search with Qdrant; — local generation with Ollama; — answers accompanied by their sources; — a refusal mechanism when the corpus does not contain enough information to answer.

My next steps focus on hybrid search, reranking, automated evaluation using recall@k, MRR and groundedness, as well as latency optimisation and cloud deployment.

After several years working in information systems, Data/BI and the operation of critical Big Data platforms, I am now strengthening my specialisation in AI Engineering through this project and regular mentoring with a Microsoft CoreAI engineer, with a focus on robustness, evaluation and production readiness. Microsoft’s CoreAI organisation is focused on building the end-to-end AI application and agent stack, so “Microsoft CoreAI engineer” is the appropriate English wording here.

I am looking for an AI Engineer opportunity focused on RAG and LLM systems, or a hybrid Data/Platform-to-AI role within a team working on real-world projects.

In particular, I can bring:

— hands-on experience with critical production environments; — a strong foundation in Data Engineering, Linux, Python, Docker and CI/CD; — an approach to AI that considers reliability, evaluation, latency and operability from the start.

To my network: if you know a team looking for this kind of profile, an introduction, a targeted recommendation or sharing this post could create a genuinely valuable connection for both sides.

I would also be happy to connect with recruiters, technical leaders and engineers working on these challenges. #AIEngineer #RAG #LLM #DataEngineering #OpenToWork

AI Digital Marketing Strategy Architect Design (Draft 1 of Project)

AI Digital Marketing Strategy Architect application. Here is a high-level overview of the project's architecture and features:

Core Functionality: The application is a comprehensive 4-phase digital marketing strategy builder that helps users plan, build, and automate their growth engines. It's designed with an elegant, editorial-style interface featuring high-contrast typography and clean layouts.

Phase 1: Foundation: A module for establishing the product's niche, generating detailed target personas using AI, and setting up the foundational tech stack (analytics, landing pages, email lists).

Phase 2: Content Engine: Focuses on organic reach. It includes an interactive SEO topic cluster generator, a visual ContentMap graph to visualize pillar links, and an AI-powered repurposing studio that transforms a single article into multiple formats (threads, LinkedIn posts, video scripts).

Phase 3: Growth & Acceleration: Built for paid and organic scale. Features a paid ad testing matrix with AI "Scale vs Kill" rules, an organic social winner analysis simulator, and an Answer Engine Optimization (AEO) readiness auditor.

Phase 4: Scale & Automation: Focuses on workflow automation. Includes a visual Zapier/n8n pipeline builder, a live lead simulator for AI-personalized email drafting, and an A/B testing loop optimizer.

Technical Architecture: A full-stack application utilizing React on the frontend and an Express server on the backend to securely manage Gemini API calls. It persists project state locally, includes preset templates (like SaaS or E-commerce), and offers a robust export engine to generate Markdown or Raw JSON blueprints.

Clean work