Freelance consultant · Senior Data Scientist · Paris, France
Causal ML and experimentation for marketing teams.
I'm Naoufal Acharki, a senior data scientist working with marketing and product teams as a freelance consultant. I design A/B tests, uplift models and recommender systems that show which actions actually change customer behaviour, and I get them running in production.
Selected outcomes
- €1M+ Incremental revenue per month Air France-KLM · A/B-tested flight-selection and offer-display recommenders · 2025–26
- Up to 30% Higher conversion in digital campaigns Senzai · uplift-based targeting across 3M customers and 50M interactions · 2023–25
- ICML 2023 Peer-reviewed research in causal machine learning Meta-learners for multi-valued treatments · plus a CSDA journal paper and a PhD from École Polytechnique
Case studies from production, not notebooks.
Personalising flight and fare offers with A/B-tested recommenders
Experimentation and evaluation for two recommender systems on the booking flow: which flights to rank first, and which fare to recommend. Every change is validated in an A/B test before it ships.
Uplift-based targeting for digital marketing campaigns
Designed and deployed the recommendation system that decides which customers to contact, when, and through which channel, using causal inference rather than propensity, and built the MLOps platform it runs on.
Causal inference and uncertainty quantification for energy production
Research at the intersection of statistical learning and causal inference: calibrated prediction intervals for gas-well production, and treatment-effect estimation for geothermal wells, leading to a patent filing and publications at ICML and in CSDA.
Evaluating the mathematical reasoning of frontier language models
Designed mathematical and statistical problems that test multi-step reasoning in state-of-the-art language models, contributing to evaluation frameworks used by leading AI research labs.
What I can do for your team.
- 01
Experimentation programmes
A/B test design, power and sample-size analysis, guardrail metrics, live monitoring and clear ship-or-stop decisions. From your first test to a running programme.
- 02
Uplift modelling and targeting
Causal models that find the customers whose behaviour a campaign actually changes, so budget goes to persuadables rather than to people who would have converted anyway.
- 03
Recommender systems and personalisation
Ranking and offer recommendation on booking, checkout and CRM flows, evaluated offline first and then validated against a control group in production.
- 04
Production ML and MLOps
Pipelines, CI/CD, model tracking and monitoring with Airflow, MLflow, Docker and Kubernetes on AWS or GCP, so the effect you measured holds after launch.
From business question to shipped result.
- 1
Start from the decision
Agree with stakeholders on the business question, the success metric and the constraints before touching the data.
- 2
Test before you trust
Offline evaluation first, then a properly powered A/B test with live monitoring. Ship what wins, stop what doesn't.
- 3
Make it run in production
Models ship with pipelines, monitoring and CI/CD, so the result measured in the experiment holds after launch.
Industry and research, side by side.
- Nov 2025 – PresentSenior Data Scientist, Marketing Operations Research · Air France-KLM
A/B testing and recommender systems for flight selection and offer display.
- Aug 2025 – Jul 2026Mathematics and Statistics Expert · Mercor
Problem design for evaluating the reasoning of frontier language models.
- Oct 2023 – Oct 2025Senior Data Scientist and Machine Learning Engineer, Pipelines and MLOps Lead · Senzai
Uplift-based targeting for digital marketing campaigns, deployed to production.
- May 2023 – Oct 2023Data Scientist, Geo and NLP team · namR
Large-scale feature engineering and pipeline optimisation on GCP.
- Oct 2019 – Dec 2022Research Engineer, AI and Data Science R&D (industrial PhD, CIFRE) · TotalEnergies One Tech
Statistical learning and causal inference for energy production.
Selected publications
- Comparison of meta-learners for estimating multi-valued treatment heterogeneous effects ICML 2023
- Statistical learning and causal inference for energy production PhD thesis, École Polytechnique, defended in November 2022
- Robust Prediction Interval estimation for Gaussian Processes by Cross-Validation method Computational Statistics and Data Analysis, 178:107597, 2023. doi 10.1016/j.csda.2022.107597
Recent talks
- Aug 2022Greek Stochastics μ' 2022: Causal Learning Talk · Corfu, Greece
- Jun 2022MASCOT-NUM 2022 annual meeting Talk · Clermont-Ferrand, France
- Jun 2022The Causal TAU seminar Invited talk · Inria · Gif-sur-Yvette, France
- May 2022Journée Causalité/XAI Talk · SINCLAIR, EDF Lab Paris-Saclay · Palaiseau, France
Working on a measurement, targeting or personalisation problem?
Available for freelance and consulting engagements in marketing analytics, experimentation and causal ML, from a single A/B test to a full personalisation programme. Project-based or longer engagements, remote or in Paris, France; I work in English and French.