Cheikh Bilal

Au sein du Lab IA, je mène des travaux de recherche appliquée sur les sujets les plus avancés du gen coding : agents de code, MCP, context engineering, sécurisation des toolchains et exploration profonde de codebases. Mon focus : concevoir les couches qui permettent à ces systèmes de réellement comprendre, parcourir et manipuler du code en environnement réel — via orchestration agentique, guardrails, proxies de sécurité, compréhension structurelle du code, AST-RAG, embeddings locaux et vector stores on-device.

Our mission: Creating AI monsters that shatter the limits of human productivity. Our intergalactic apps turn days into minutes. Our services transform your ideas into apps.

Accompagnement des équipes dans la création d'applications web (Front-end et Back-end) dans le secteur de l'investissement. Stack: Cloud: Azure Backend: Python, Fast API, MS SQL, SQL Alchemy, Docker Front-End: React

The Antares team is developing an open source web platform to simulate the supply-demand balance and measure the adequacy or economic performance of interconnected electrical systems, both in the short and long term. Initially created in the form of software to download, the objective of my mission was to help the team to create a Web version (Web app), to redesign the product to make it much easier to use for the different teams, with a simple and ergonomic design. Within this team, I was one of the main developers, and involved both in the back-end part (Python, Fast API) and responsible for the Front-end part (React, Typescript, Material UI).

Development of software tools to simulate and test the radio base stations of new 5G infrastructure. Belonging to a "Tools" team working in Scrum Agile, I have a lot of autonomy in the technical choice of our products. I am involved in all phases of the projects I develop, and also act as technical referent. I bring my contribution in the software architecture, development, integration on target as well as in the training of the various teams (testers and integrators). Technical environment: Language: Python, C++ Compilation tools: CMake Environment: Linux Tools: Jenkins, Git, Jira Management: Agile

Moonster is a start-up developing a whole new social file sharing system in the Cloud. The idea is to allow you to create a virtual hard drive shared between different groups of friends or colleagues. As a technical lead, my tasks are essentially of two types: 1) Development of a functional prototype 2) Team management Technologies: Client side: React.js (web) Back-end: Node.js / Express.js / REST API Database : MongoDB

Inside the Enterprise & Cybersecurity department, my mission at Gemalto was to develop C# software to reprogram smart cards, for demonstration purposes to future customers. Intended for non-technical teams, the software had to be user-friendly and offer a responsive and ergonomic interface. Within an Agile team of 8 people, in an ultra-sensitive and demanding context, I managed each phase of the project : - Listen to the requests of my customers (sales teams) - Produce specifications - Make the right technical choices - Build the software architecture - Develop the software - Improve the software based on feedback

As part of the international DUNE experiment (Deep Underground Neutrino Experiment), within the Department of Particle Physics at the University of Oxford, my work consisted in verifying certain hypotheses concerning Neutrino particles, by developing simulations. These Neutrino particles, still very mysterious, could explain the dissymmetry between matter and antimatter in the observable Universe. But to be able to study them, we must already succeed in detecting them. The DUNE experiment provide a couple of underground particle detectors, equipped with a chamber filled with liquid argon at low temperature, and electrodes. These detectors will detect the Neutrinos that roam the earth and reach us from the rest of the universe. To verify the hypotheses of the DUNE experiment, I was able to take part in the development of a computer simulation of one of the two detectors. This simulator was developed in C / C++ language, using frameworks developed by CERN (Roots and Geant4) and dedicated to research in particle physics. The simulations turned out to be successful, since theoretical assumptions regarding neutrino detection were verified. Hoping that the real experience will also be a success and advance our understanding of the universe.