Who I am
I am a French citizen based in Saint-Julien-en-Genevois, France, just outside Geneva.
For more than ten years I have worked with international organisations, in multicultural teams and delivering in many countries. My work has focused on helping businesses grow and create impact.
I studied Economics in Sorbonne University Paris, which shaped how I read markets, incentives and the drivers of competitiveness. More recently, I pursued a Master's programme in AI and Digital Technology at ETH Zurich, adding technical depth to that economic and business foundation.
What drives me
What drives me is understanding what an organisation wants to achieve, what is holding it back and what can help it move forward.
My curiosity first led me to data, which I used to spot business opportunities and support better decisions. I then moved into Business Analysis, turning data and business needs into practical digital solutions.
Today, AI is the next step in that journey. I help companies make the most of it by choosing the right technologies and use cases, implementing them responsibly and staying compliant with regulations. But I see AI as more than another tool added to existing structure. It reshapes how companies operate, requiring them to rethink their strategy, processes and organisation.
A closer look at my experience
This section complements my CV by providing further context and detail on the skills. It presents five dimensions of my experience and shows how they come together in the responsibilities I take on.
I coordinate projects from initial scoping through deployment and handover. I coordinate resources, teams and partners while balancing project objectives, budget, timelines and operational constraints.
- Defining project scope, objectives, deliverables and implementation strategy
- Developing work plans, timelines, responsibilities and dependencies
- Managing project budgets and preparing inputs for management reporting
- Recruiting, allocating work to and supervising consultants
- Drafting terms of reference, assessing bids and recommending suppliers
- Coordinating developers, service providers and institutional partners
- Monitoring operational risks, timelines, quality and deliverables
- Preparing deployment, training and handover
Examples of key decisions
- Reducing the functional scope to match the technical and financial capacity of future users
- Prioritising long-term sustainability and user adoption over technical sophistication
- Prioritise outsourcing part of the development work to optimise costs
- Reallocating part of the budget to training rather than additional features
- Bringing the implementation schedule forward to anticipate institutional changes
I go beyond the stated request to understand the underlying business need, the processes involved and the expected value. I translate this understanding into target-state processes, functional requirements and clear priorities for business and technical teams.
- Analysing current processes and designing target-state processes
- Mapping processes, workflows and data flows
- Eliciting, analysing and clarifying requirements
- Identifying and assessing use cases
- Defining functional architecture
- Writing functional requirements and specifications
- Designing prototypes and wireframes
- Prioritising features
- Bridging business and technical teams
- Conducting functional testing and identifying gaps
Examples of key decisions
- Reconciling changing or conflicting needs to maintain coherent specifications
- Prioritising essential features according to their value, cost and users’ capabilities
- Embedding human validation at critical points in the process rather than pursuing end-to-end automation
- Requiring rework when deliverables do not meet agreed requirements or user needs
I understand the mathematical and statistical foundations that underpin models, as well as their limitations. This enables me to select methods that fit the business need, interpret and anticipate model behaviour, and assess the associated benefits and risks.
I incorporate AI governance from the design and evaluation stages, covering data privacy, re-identification risk, bias and fairness, transparency, acceptance criteria, documented trade-offs and deployment conditions. I also assess how relevant GDPR and EU AI Act requirements affect the design, evaluation and use of AI solutions.
I can design and implement the following technical methods and evaluation frameworks in Python. For production engineering, industrialisation and end-to-end deployment, I work with technical specialists.
- Data preparation, exploratory data analysis and feature engineering
- Training, tuning and comparing machine-learning models
- Linear and logistic regression, Random Forest, XGBoost and neural networks
- Designing strategies to address class imbalance
- Generating synthetic data with CTGAN, DPCTGAN and PrivBayes
- Designing evaluation frameworks covering fidelity, data privacy, utility and fairness
- Assessing data privacy and re-identification risk
- Auditing bias and fairness across demographic groups, including subgroup fraud rates and TPR/FPR disparities
- Defining use-case-specific acceptance criteria and thresholds
- Considering relevant GDPR and EU AI Act requirements when defining evaluation criteria, documentation and deployment conditions
- Translating business, risk and compliance expectations into evaluation criteria
- Documenting model limitations, risks, trade-offs, mitigation measures and conditions of use
- Evaluating model performance using appropriate metrics, including precision, recall, F1-score, specificity and ROC-AUC
- Designing prompts and evaluating generative-AI agents against defined business rules
- Developing in Python with Pandas, NumPy, scikit-learn, PyTorch, TensorFlow, SDV and SmartNoise
- Working with Jupyter Notebook, Google Colab and GitHub
Examples of key decisions
- Structuring model evaluation around four complementary dimensions — fidelity, data privacy, utility and fairness — with acceptance thresholds defined in advance
- Addressing class imbalance through minority-class oversampling while retaining the full majority-class distribution
- Evaluating models against multiple performance metrics rather than relying on accuracy alone
- Retaining sensitive demographic attributes so that bias can be measured and fairness audited
- Tailoring usage recommendations and deployment conditions to the model’s risk profile and applicable regulatory requirements
- Withholding a deployment recommendation when an otherwise high-performing model reveals significant fairness disparities between demographic groups
I analyse complex data to surface the insights that matter most for decision-making. I go beyond statistical results by placing them in context, assessing their implications and translating them into recommendations that decision-makers can act on.
- Checking data quality and consistency before analysis
- Conducting advanced statistical analysis and visualising results
- Analysing enterprise surveys across companies of different sizes and sectors
- Analysing large international-trade databases
- Identifying trends, constraints, root causes, costs and impacts
- Identifying priority products, promising markets and market-positioning opportunities
- Contextualising findings based on economic and institutional realities
- Developing strategic and operational recommendations
- Producing technical reports and presenting findings to decision-makers
- Working with Excel, Trade Map, Market Access Map and Export Potential Map
Examples of key decisions
- Tailoring survey methodologies to national contexts while maintaining cross-country comparability
- Strengthening data-quality checks before the data is handed over to statisticians
- Combining statistical evidence, qualitative feedback from companies and country context to avoid recommendations disconnected from local realities
- Taking the analysis beyond headline findings to connect observed barriers with their root causes, costs and impacts
- Turning data insights into actionable recommendations and concrete projects
I design and deliver programmes that enable participants to understand, adopt and independently use new tools, data and methods. In international and multicultural environments, I tailor my approach to local needs and build engagement through listening, demonstration, negotiation and consensus-building.
- Analysing training needs and defining learning objectives
- Designing training materials, case studies, exercises, tests and evaluation questionnaires
- Delivering training, workshops and advisory assignments in French and English
- Training public-sector decision-makers, trade-support institutions, business associations and companies
- Delivering assignments across at least 15 countries
- Translating technical and analytical topics into accessible and directly applicable content
- Supporting users through the adoption of new tools and processes
- Managing resistance, expectations and varying levels of technical maturity
- Mobilising stakeholders without direct hierarchical authority
- Communicating and collaborating effectively across cultures
Examples of key decisions
- Involving future users throughout development to support adoption and effective handover
- Adapting the pace, content and level of training to participants’ technical capabilities
- Delivering additional sessions where necessary to build lasting user autonomy
- Extending training beyond tool use to cover long-term sustainability, future ownership and often-overlooked operating costs
- Addressing resistance by demonstrating value, negotiating and building consensus around workable solutions
Looking ahead
- I am strengthening my command of Agile methods and Product Ownership to better structure my delivery practices, manage the backlog and guide the evolution of solutions beyond launch.
- I am also continuing my technical work in machine learning. As neural networks are often seen as “black boxes”, I am working on a methodology to understand the explainability of some deep-learning models. The aim is to better understand what drives their predictions and find the right balance between performance and transparency, in sectors where decisions need to be explainable and verifiable.
- I also keep up with developments relating to the GDPR, the EU AI Act and other AI governance frameworks, as they directly influence use-case selection, the use of data, and deployment conditions.
I aim to take on increasing responsibility for leading Data & AI projects, combining strategic vision, strong technical understanding and responsible AI governance.
Let's connect
Would you like to discuss a need, a project or a professional opportunity? I would be happy to explore how I could contribute.