Behavioral AI
Implicit signal modeling, predictive segmentation, personalization and behavioral understanding.
Based in France · Open to remote opportunities
PhD in Cognitive Science · Research Engineer
I turn human behavior
into intelligent products.
5+ years designing and deploying scalable AI systems for high-traffic products, using research-driven methods across user behavior, data and UX.
100M+Multimodal datapoints analyzed
5+Machine Learning POCs delivered
3ML products shipped
5,900+Scientific publication reads
01 — About
From data and research to product.
I use research as a practical tool to guide product decisions — from hypothesis formulation and experimentation to production deployment.
My projects turn complex behavioral, visual and semantic signals into usable product insights. The focus is on fast iteration, interpretability and real-world impact.
02 — Expertise
A full-stack approach, from people to models.
Search & Recommendation
Semantic search, ranking, embeddings, visual similarity and recommendation systems.
Quantitative UX Research
Experimental design, eye tracking, mouse tracking, behavioral metrics and causality.
Applied ML Engineering
Robust prototypes and ML services with Python, FastAPI, PyTorch, scikit-learn, Docker and GCP.
03 — Selected work
Research rigor. Tangible impact.
01
↗02
↗03
↗04
↗05
↗06
↗
Flagship · Computer Vision
AttentionAI — visual attention prediction
Predict attention areas from a single interface image to optimize designs before production.
Visual Attention · Transformers · TensorFlow · Vertex AI↗02Multimodal Search
SearchAI — visual and semantic search
Explore visual catalogs through text-image embeddings, from ingestion to scalable vector retrieval.
CLIP · Vector Search · Cloud Run · FastAPI↗03Behavioral ML
CursorAI — early engagement prediction
Predict engagement in the first seconds of a session from mouse trajectories, enabling real-time personalization.
Mouse trajectories · XGBoost · Real-time inference↗04NLP · Psychometrics
TalentMap — personality and soft skills
Infer personality and soft skills from free text, retrieve similar profiles and map an entire candidate pool.
NLP Embeddings · Psychometrics · Vector Search↗05UX Prediction
DeepUX — automated UX evaluation
Score UX dimensions from website screenshots with deep learning, turning qualitative principles into measurable signals.
CNN · Behavioral ML · Scikit-learn · FastAPI↗06Synthetic Research · UX Intelligence
Synthetic UX — simulated product research
Frame a study, generate a synthetic cohort and connect simulated answers, behavioral profiles and visual signals in one workspace.
ClickHouse · Embeddings · LLM · FastAPI · Next.js↗Scientific publication
SCIENTIFIC REPORTS · NATURE PORTFOLIO · 2024The way we look can reveal part of our personality.
Introducing gaze-based signatures of personality through image and webpage exploration experiments.
View my publications ↗04 — Journey
Across research, product and engineering.
Research Engineer · AI R&D
AB TastyPythonGCPVertex AIDockerLLMs
Applied behavioral research and AI systems for high-traffic digital products.
- Behavioral scoring deployed at the scale of millions of users
- Experimental protocols and actionable insight delivery
- Product, UX and commercial decisions with product and leadership teams
Research Engineer · PhD Research Scientist
Dotaki / AB Tasty · Université Paris CitéEye trackingMouse trackingMachine LearningPsychometrics
Personality, visual attention and user interactions studied through gaze and mouse-movement signals.
- Contribution to EmotionsAI, real-time behavioral segmentation based on emotional needs and mouse movements
- 100M+ eye- and mouse-tracking datapoints analyzed
- Behavioral metrics integrated into ML systems and a Scientific Reports publication
UX Researcher · Data Science Intern
ContentSquare · Sublime Skinz · StimCo · Neuroscience LabPythonUX ResearchData ScienceCNN
Early work across UX research, cognitive psychology and Deep Learning.
- Quantitative UX studies and data analysis
- Eye tracking applied to visual attention
- Convolutional networks for image classification
PhD & dual master’s background
Université Paris Cité · Grenoble INPCognitive ScienceCognitive PsychologyResearch
PhD in Cognitive Science, MSc Cognitive Science and MSc Cognitive Psychology.
- Experimental design and quantitative methods
- User-behavior modeling
- Cognitive-science engineering
STACKPython · FastAPI · scikit-learn · PyTorch · TensorFlow · XGBoost · Docker · GCP · ClickHouse · Meilisearch · SBERT · CLIP
05 — Contact
Let’s build products that truly understand their users.
I’m open to roles in Behavioral AI, Applied ML, Quantitative UX Research and Product Data Science.
Start a conversation ↗