I make AI agents work in your team, reliably.

Deltarm

Senior fullstack developer · AI agents

In a team of six developers, I build a governance, risk and compliance (GRC) platform for CAC 40 groups: risk register, internal control, action plans, data collection, reporting and mapping.

The team openly uses AI intensively: for each ticket, agents implement, review and test all the way to delivery, under technical and product supervision. In six months, it delivered what a conventional team of the same size would have needed more than three years to build (a conservative estimate, based on the features delivered), without cutting corners on tests.

I contribute to:

• CI reliability: flaky-test quarantine and tracking, E2E tests refocused on user journeys;
• the agents' skills: PR review, fixing review findings, hunting flaky tests;
• the agents' working rules: review by an AI, human acceptance testing, no auto-merge;
• business features: mapping, data collection, internal control, action plans, reporting.

SecOps Forces

Fullstack developer · sole developer of the application

I build the SaaS platform of SecOps Forces, which protects SMEs and mid-size companies: email and domain security (DMARC, SPF, DKIM, leaked credentials, certificates), phishing and awareness campaigns, monitoring of vulnerabilities and lookalike domains, and continuous Microsoft 365 audits against the CIS and CISA benchmarks.

I added an AI assistant, based on Mistral, that answers from each organisation's security data.

On the build side:

• front office and back office in React, a multi-tenant .NET API, Stripe billing;
• more than 3,000 automated tests, from unit tests to Playwright user journeys;
• AI-agent-assisted development, with project-specific conventions and skills.

DeepDetective

Fullstack & AI developer

For an investigation board game with a companion app, I designed the AI that lets players talk with the characters: interrogations become real conversations, and the game comes alive.

As the sole developer, I also took over and hardened the existing app: data security, player progress, performance.

On the AI side:

• it replaces a third-party chatbot service;
• it queries several LLM providers, with automatic fallback;
• it draws on the scenario through RAG, and streams its answers;
• the game's writers author the characters in Notion, and they are synced automatically.

Ynov Campus · IPSSI

Master's teacher

Passing on knowledge has always driven me. I have taught at Ynov Campus Sophia-Antipolis since 2022, and I taught at IPSSI in 2025, in computer science Master's classes: AI-agent-driven development, AI integration, fullstack web development, software testing, algorithms and Web3.

Smart-Chain

Fullstack developer, Web3 & AI

A Paris startup in the blockchain and artificial intelligence space, where I developed several products, including one for the French Alternative Energies and Atomic Energy Commission (CEA), among other prestigious projects.

Fullstack developer on a cloud infrastructure, mainly using NodeJS, TypeScript, React, Next, Prisma, PostgreSQL, WebSocket, and queuing, to build web applications interacting with:

• blockchains: smart contract calls, indexing, IPFS;
• AI models I developed in-house in computer vision (CNN, GAN, ViTransformers) and on tabular data (RNN, LSTM): classification, detection, segmentation, and regression;
• LLMs such as those from OpenAI and more recently Mistral AI: assistants, prompting, fine-tuning, and RAG.

Ricardo

Frontend, then backend developer

I worked on the first marketplace in Switzerland, in an English-speaking company.

One year as a frontend developer with React hooks and TypeScript based on hydration, using Material-UI, React-Query, and Redux.

Then one year as a backend developer with Java Spring, .NET, Kafka, PostgreSQL, and SQL Server, on a Kubernetes infrastructure managing our microservices, with Scrum.

I also improved my English through daily interactions with my English-speaking colleagues and various presentations.