# Simon Potel — Full-stack software engineer & AI Engineer > Full-stack software engineer & AI Engineer based in Lille. Custom development, project architecture and AI integration, for real business impact. - Location: Lille, France - Email: simon@potel.dev - LinkedIn: https://linkedin.com/in/simonpotel - GitHub: https://github.com/simonpotel - Website: https://simonpotel.com/en (Français: https://simonpotel.com/fr/index.md) ## About I'm a developer, and what I love most is **solving problems**. Today, I put my development expertise to work for real business impact: laying out a project's architecture, building it end to end, and augmenting it with AI where it brings a real gain. At home in the major cloud, data and AI ecosystems: I adapt to whichever one the project calls for, and I know how to get the most out of each. Ecosystems: Anthropic, OpenAI, Google Cloud, AWS, Cloudflare, Microsoft Fabric. ## Experience Four roles at Advanced Schema since 2025: from data-flow monitoring to AI agents. ### Advanced Schema — Data-focused IT consultancy, Lille #### Full-stack developer · Work-study · Oct 2026 – present Between the Dev Factory and the AI Factory, I evolve the company's products and design new solutions for clients in luxury, banking and retail. - **AI-augmented products**: evolving existing products and launching new ones, with AI where it cuts costs or solves a complex problem. - **AI agents in production**: monitoring, diagnosing and fixing production, speeding up development, always under human approval. Tech stack: TypeScript, Python, NestJS, Nuxt, LLM, MCP, Langfuse, Kubernetes. #### Full-stack developer · Internship · Jun 2026 – Oct 2026 Four months between the Dev Factory and the AI Factory, for clients in banking, luxury and retail: an agentic platform put into production, a self-remediation agent and a Cognos to Power BI migration. - **Agentic platform in production**: a consultant describes their know-how and gets a deployable agent, reusable across the company. - **Monitoring screens generated by an agent**: a few hours instead of several days per screen. - **Self-remediation agent**: an incident handled in 7 to 20 minutes instead of about 6 hours; in testing, 60% of anomalies resolved with no human intervention. - **Security and governance**: SSO, isolation between organisations, human approval of writes, a budget and traceability for every run. - **Cognos to Power BI migration**: real financial models imported with their lineage, from database to report, under more than 460 automated tests. Tech stack: TypeScript, Python, Go, C++, NestJS, Nuxt, LLM, MCP, LiteLLM, Langfuse, Keycloak, Power BI, PostgreSQL, Docker, Kubernetes, Helm, GitLab CI. #### Full-stack developer · Part-time internship · Sep 2025 – Jun 2026 Alongside my studies, a year within the Dev Factory on three fronts: a retail client's production, the company's data product and AI R&D. - **Real-time monitoring (retail)**: flow-control interfaces made clearly faster and more responsive. - **Cloud infrastructure**: secured access, deployed services, hands-on during production releases. - **Data product**: transforming and enriching data inside the governance and modelling tool the company publishes. - **AI and machine-learning R&D**: retrieval-augmented generation (RAG), vector databases, predicting relationships between data. Tech stack: TypeScript, Python, Go, NestJS, Vue.js 3, BigQuery, Pub/Sub, Google Cloud, Terraform, Docker, GitLab CI. #### Full-stack developer · Internship · Jun 2025 – Sep 2025 Within the Dev Factory, for a retail player: tools to monitor, trace and fix its financial data flows in real time. - **Real-time monitoring**: a per-flow configurable interface (volumes, success rates, gaps, alerts), where a new source plugs in within minutes. - **Data lineage**: end-to-end tracing of an object, to spot blockages and rejections at a glance. - A secure **data-correction API**: the client's partners fix their own data. - **R&D**: a ClickHouse, PostgreSQL and BigQuery benchmark (performance and cost), and a prototype AI assistant that explains a rejection in plain language. Tech stack: TypeScript, Python, Go, NestJS, Vue.js 3, REST API, BigQuery, Pub/Sub, Google Cloud, Terraform, Docker, GitLab CI. ## Expertise My goal: create value for the client, with the solution that best fits the need. I use AI where it cuts their costs, and where these new technologies make it possible to solve complex problems. ### Building your product From idea to a **reliable product in your users' hands**: I design and build your **web application**, then keep it evolving with them. - **Web applications** designed for your users - **Regular releases**, tested before each launch - A product that **grows with your business** ### Scoping and foundations The **right choices from day one**: a solution sized for **your need and your budget**, secure, and ready to grow. - A **clear scoping** of needs and priorities - Technical choices you can justify, **at the right cost** - A **secure** solution, ready to **scale** ### AI that works for your teams **Fewer repetitive tasks**, **faster decisions**: AI where it saves time or money, always **under human control**. - Assistants that answer from **your documents and data** - Repetitive tasks **automated**, approved by your teams - **Controlled costs** and a **measured gain** ## Stack To me, technology is a means to solve a problem, never an end in itself. I rely on agility and real firepower in modelling and low-level work: foundations that make me an expert with these tools, without ever being limited to them. - Languages: TypeScript, C++, Python, JavaScript, Go, C, PHP, SQL, GLSL - Frameworks: NestJS, React, Vue.js 3, Next.js - Frontend: Three.js, WebGL, WebGPU, WebXR, WebSockets, GSAP, Framer Motion, HTML5, CSS3 - Runtime & APIs: Node.js - Cloud & infra: Jenkins, Portainer, WireGuard, Raspberry Pi - Data & BI: BigQuery, MySQL, ClickHouse, Pandas, NumPy, dbt, Power BI, Cognos, DataStage - DevOps & CI/CD: Docker, GitHub Actions, GitLab CI, Git, Linux - Testing & quality: GoogleTest, Jest, unittest, TDD - AI: Langfuse, LiteLLM - Design & tools: Figma, Blender, Jira - Methods: REST API, Scrum, Agile, Microservices ## Education - SUPINFO — Computer engineering degree (5 years), Lille, 2024 – 2029 ### Licenses and certifications - AWS Academy Graduate – Cloud Foundations — Amazon Web Services, Apr 2026 - CCNA v1: Introduction to Networks — Cisco, Dec 2025