A new project to define
Start with your needs and data to design a new tool, add a product feature or simplify time-consuming work.
Common situations
When I can help
An idea to put into practice
Choose a method, develop a first version and test what it brings before committing to the next stage.
A team that needs extra capacity
I take on development or analysis, solve scientific problems and work alongside your team without a new hire.
Engagements for healthtech, medtech and biotech companies, R&D teams, consultancies, training organizations and private healthcare providers.
Ways to work together
A clear scope and a useful deliverable
We agree on the intended result, work and budget before starting. I can take on a defined assignment or reserve days to work with your team regularly.
Find the right idea
Define a useful project
Quote based on scope
A short assignment, with the scope agreed together
Identify where AI or data analysis can help your products and teams.
- Business needs and available data
- Ideas for tools or product features
- First implementation to test and expected benefit
- Work plan and budget estimate
Alongside your team
Scientific R&D support
Quote based on scope
A focused assignment or reserved days each month
Move your projects forward with extra scientific and development capacity.
- Method and prototype development
- Data analysis and modelling
- Problem-solving, error analysis and validation
- Code, results and documentation handed over to your team
A working first version
Prototype / proof of concept
Typical budget: €8,000–€15,000
Typically 4–8 weeks
Build a first version of your tool and measure its value for a specific use case.
- Reference method and success criteria
- Reproducible development and analyses
- Testing and improvements within an agreed scope
- Prototype, documentation and recommendations
Focused feasibility review
AI & data diagnostic
Decide whether a specific idea merits investment: data review, technical options and a decision memo. Typically 1–2 weeks.
Discuss feasibilityFocused evaluation
Technical model audit
An independent review of an existing model: protocol, robustness, bias, reproducibility and recommendations. Typically 2–4 weeks.
Discuss a focused auditKnowledge transfer
AI / LLM training for healthcare
Use cases, prompt design, limitations, validation, confidentiality and best practices adapted to professional roles.
Continuity
Monthly scientific advisory
The entry package covers up to one expert day per month, delivered as one workshop or two half-days.
- Monthly steering session and experiment prioritization
- Review of two short items: protocol, results or decision memo
- Decision summary and focused asynchronous exchanges within the agreed allowance
Not included: production development or analysis, continuous availability, clinical validation, certification or regulatory advice. The quote fixes the volume, cadence, deliverables and response times.
Scope a monthly engagement
From €1,500 / month
Exact scope confirmed in the quote
Indicative budgets, adjusted to scope, data quality and security constraints. A precise quote is provided before every engagement.
Method
A simple, traceable collaboration
Discuss needs and define the work
Discuss your needs, tools to build or improve, available data and the objectives of the work.
Proposal and quote
Duration or fixed scope, exact price, number of iterations, deliverables, validation criteria and payment terms.
Delivery and handover
Progress reviews, validated deliverables, documentation, final debrief and transfer to your teams.
Public expertise and methods
Research and code you can examine
My publications and their code document my methods for design, development, analysis and evaluation.
Multi-dataset predictive protocol with public code
Analysis of four public datasets using penalized regression, 10-fold cross-validation repeated 100 times and open code so that the analysis can be checked.
What this documents: protocol design, multi-dataset comparison, out-of-sample evaluation and reproducibility. It is neither a client case nor clinical validation.Quantitatively evaluated image pipeline
A specialized image generation, encoding and editing pipeline, with quality and editing-level measures derived from real-data distributions.
View the ACM publication What this documents: complex pipeline construction, metric design and constraint checking. It is neither a client case nor validation of a healthcare model.Validating a predictive model in healthcare
Population, outcome, data leakage, dataset splitting, metrics, calibration, subgroup analysis, reproducibility and final decision.
Lire le guide de validation What this provides: an operational checklist that turns these principles into explicit audit criteria adapted to the project context.Possible scopes
Types of topics that can be addressed
These examples illustrate possible scopes. They are not published client cases.
- Automated quality control for biomedical images: framing, artifacts and noisy data.
- Completeness checks for administrative or research records using tabular data, without autonomous clinical decisions.
- Error analysis and robustness improvements for a predictive model used internally.
- Structured extraction of variables from medical or scientific PDF documents.
- Scientific dashboard for exploring cohorts, indicators or model results.
- LLM training for healthcare managers: useful applications, prompt design, verification, confidentiality and limitations.
Working framework
Security, validation and explicit limitations
Controlled data
Work on data that are genuinely anonymised where feasible, on synthetic data, or directly within the client’s controlled environment under the applicable legal and security framework.
Documented validation
Baselines, data splits, appropriate metrics, error analysis and limitations recorded in the deliverables.
Responsible use
Prototypes support analysis and decision-making. They are neither certified medical devices nor clinical advice.
This consulting practice is independent and does not represent École Normale Supérieure or INSERM.
Next step
Let’s discuss your needs.
Whether you have an idea, an ongoing project or a need to clarify, a few lines about your work are enough to start the conversation.