As co-founder of Stackeo, I developed a B2B SaaS solution for decision makers in IoT. I was mainly in charge of leading the Product Discovery at crossroads with Strategy, Business and Customer Success challenges. My main tasks were:- leading Product Market Fit research, - driving user research & building Ideal Customer Profile, - prototyping MVP (Minimum Viable Product) and patenting features, - delivering in app using agile development practice in CI/CD.When co-managing a startup, I had to build a Business Plan, investigate Market, Set priorities and driving by impact, Hire talented coworkers, work on fundraising, convince in business meetings, participate in showcases and events.Company pitch:Stackeo is a SaaS company helping enterprises to industrialize their connectivity and IoT (Internet of Things) solutions at scale. We provide a suite of tools to articulate IoT strategy, align IT/OT teams and pilot sustainable value chains. Stackeo develops the concept of IoT-Architecture-as-Code, based on StackML, a dedicated modeling language and a patented systemic methodology.

Lead various projects of development of B2B SaaS applications for strategic decision making for large industrial companies.- Bring technical support in sales meetings- Lead product development (MVP)- Lead workshops with C-level clients- Investigate product market fit for every MVP- Code complex system simulations and ML algorithms (C++, python)Some relevant missions detailsClient: NexansRole: Project ManagerSubject: Product Market Fit of an Asset Management SaaS solutionKey results: Built strong partnership with CAC40 company. Reached the EU DSO market. How? Conduct user research (buyer persona and ICP), Test value proposition, Make product demo and sales material kit, Build trust with SME decision makersClient: SolvayRole: Product managerSubject: Reduce downtime in Oil&Gas industryKey results: Delivered a SaaS MVP combining a data pipeline (ETL), a predictive simulation tool and BI dashboards (plant OEE)How? Work on site with Process Plant Management team, Understand industrial challenges and convert it into product needs, Deep dive in the process science, Lead product developmentClient: Veolia - SedifRole: AI expert developerSubject: Crisis Management for Water supply network distributing 10M people "ServO"Key results: Delivered a product feature for finding optimal operation planHow? Conducted functional analysis, Created ad-hoc multi-criteria optimization tool on graph topology, Implemented custom Genetic Algorithm and Ant colony Algorithm (C++), Test at-scale on real datasetsClient: Petrogal, Dow Chemicals, Fertinagro - TOP-REFRole: Project managerSubject: Increase process performance in Oil&Gas and Chemicals industryKey results: On-site deployment of a Machine Learning solution for process optimization, Proven results published, Define EU standardsHow? Benchmark of ML tools, Built Surrogate model using Kriging (Python, Scikit-learn),

Food process design is a complex activity facing the wide diversity of products and processes of the food industry. In southern countries, the dramatic rise of demand drives the need of improving existing traditional processes at industrial level. In this thesis, it was attempted to develop some rational mathematical method to deal with the multiplicity of objectives that have to be simultaneously satisfied when designing some food process. It was applied to the specific case study of the hot-smoking process for the production of traditional smoked catfish in an African context. First, a model was built and proved to be able to predict the process behavior with coherency. Then, the global optimization method enables to find very efficient and satisfactory design solutions. This original work raised the difficulties of the design activity in the food industry sector, especially the lack of relevant expert data.

Project of developping new devices for rapid thermal caracterization of PCM, which has growing domains of application in thermal energy storage for CSP (Concentrating Solar Plants) or Buildings energy savings (HQE Buildings in France). Studies of long-term behaviours' caracterization, and High temperatures modelling.Studies on undercooling characterization with thermal camera.

Developed post-processing of meteorological data of Large Eddy Simulation Model (LES) in a Linux environment.Studied fluid mechanics of both boundary layer meteorology and turbulenceWorked with Canadian Numerical Weather Prediction models

Optimized an antenna according process of a mass production device (potential benefit ~ 40k€ per year)Took part in negotiation with foreign suppliersNB : MGP Instrument is the global leading company of both Dosimetry & Radiation monitoring systems ; Part of Mirion's Group.