Vincent B.

AI Engineer specializing in LLMs and Data, I design useful AI systems in real-world contexts: business assistants, LLM workflows, automations, RAG, documentary or analytical copilots, content summarization, and decision support. My approach connects business framing, industrial rigor, and technical execution, with a simple objective: transform an operational need into a testable prototype, then into an exploitable deliverable.

Since 2019, I have been supporting AI/Data projects involving Python, SQL, REST APIs, FastAPI, Docker, LangChain, LlamaIndex, embeddings, multi-model evaluation, and orchestration. I work with LLMs and tools like OpenAI, Claude, Gemini, DeepSeek, and local LLaMA when confidentiality is critical, as well as with automation and delivery environments such as n8n, VS Code, OpenAI's Codex, or OpenClaw depending on the context. My use cases cover customer service assistants, internal support, web summarization, content generation, AI workflows, dashboards, classification, segmentation, prediction, and decision support.

Before AI, I built solid experience in demanding industrial environments—Safran, Renault, PSA, Nexter—and in data-driven management at Scalian for Valeo, involving cross-functional coordination, reporting, and execution under constraints. This background enables me to deliver lean, measurable solutions that are understandable by business teams and technically defensible.

I am involved in framing, prototyping, and stabilizing Applied AI / GenAI / LLM / Automation solutions that quickly create value without losing sight of reliability, confidentiality, and user adoption.

Before August 2019, I built a solid foundation as an industrial engineer in the automotive, aeronautics, space, defense, and energy sectors. My roles gradually shifted my focus from traditional mechanical engineering to a more data-oriented approach: numerical simulation, correlation of calculations/measurements, signal processing, structuring test data, automating post-processing, visualizing results, KPIs, reporting, and dashboards.

At PSA, Renault, Nexter, APCO, and Safran, I worked on problems where technical data was central: validating models, comparing results, exploiting measurements, reducing calculation times, and transforming complex analyses into engineering decisions. At Valeo, this evolution was reinforced with QCDP (Quality, Cost, Delivery, Planning) management, KPIs, quality/cost/time analysis, and cross-functional project monitoring.

This background explains the coherence of my current positioning: I am not approaching Data and AI as a trend, but as a natural continuation of a career built on measurement, modeling, automation, reporting, and decision support.