Data science · sports & health apps · automation · agentic AI · data engineering

From data to apps: models, pipelines, and products for sports, health, and research.

I help teams end to end — data scoping, experimentation, lightweight MLOps, field mobile apps, and dashboards — for measurable decisions and real-world use.

Discuss a project View interactive demos
Maxime Rosito

Focus areas

Three ways I can help

The same bar: clean datasets, interpretable models, software that ships and lasts — from notebooks to production apps.

Data science & data engineering

Ingestion and prep, features, supervised or time-series models, rigorous evaluation, monitoring — from analysis notebooks to schedulable pipelines, plus automation (CI, orchestration) and agentic AI delivery (tools, guardrails, observability).

Sports, health & field apps

Design and development of mobile apps for athlete monitoring, training, or wellness: sensor sync, staff workflows, UX, and store release.

Science, signal & research

Numerical prototyping, signal processing, reproducible experiments — MATLAB, Python, tying literature to real-world constraints.

Expertise

Technical breadth

A hybrid data scientist / engineer profile — focused on data products and sports & health apps, with automation and agentic AI where they add value.

Data science & pipelines

Python, pandas, scikit-learn, SQL, ETL, notebooks, visualization, tests & lightweight MLOps

Sports / health & mobile

Android, Kotlin, Java, APIs, data & sensor sync, real time, release & store

Science & computing

MATLAB, modeling, signal, experimental data & research protocols

Web platforms

TypeScript, React / Vue, Node, APIs, dashboards & pragmatic deployment

Automation & agentic AI

CI/CD, job orchestration, LLM integration, agents & tools (RAG, function calling), guardrails, tracing & cost control

Approach

Rigorous, without needless jargon

  • Data framing — business question, metrics, data quality (sports, health, lab), GDPR and confidentiality.
  • Short iterations — testable analyses or features, app demos or interim reports with staff and users.
  • Handover — documented notebooks or pipelines, model performance criteria, transfer to your data or product teams.
Publications

Scientific work

15 references · click to show or hide the list

Full bibliography on Google Scholar for Maxime Rosito — articles, journal abstracts, posters, and talks; DOIs are listed when available (ORCID).

  • Article

    Cortical processing for the vestibular and visual input of egomotion in macaque monkeys: Separate networks with targeted convergence

    Imaging Neuroscience 4, IMAG. a. 1093 · 2026

    Article (DOI) bioRxiv preprint (2025) Google Scholar

  • Journal abstract

    Perceptual Learning as a Rehabilitation Approach to Enhance Motion Processing in Maculopathy Patients

    Investigative Ophthalmology & Visual Science 67 (3), 52-52 · 2026

    Google Scholar

  • Article

    Visuo-vestibular integration for self-motion: Human cortical area V6 prefers forward and congruent stimuli

    Experimental Brain Research 243 (6), 152 · 2025

    Article (DOI) bioRxiv preprint (2024) Google Scholar

  • Conference · poster

    Perceptual learning improves motion perception in patients with macular degeneration

    46th European Conference on Visual Perception (ECVP), 2024

    Google Scholar

  • Talk

    Perceptual learning improves motion perception in patients with age-related macular degeneration

    Annual GDR Vision forum (GDR Vision 2024) · 2024

    Google Scholar

  • Article

    Time-to-contact perception in the brain

    Journal of Neuroscience Research 99 (2), 455-466 · 2021

    Google Scholar

  • Article

    Key periods of cognitive decline in a nonhuman primate model of cognitive aging, the common marmoset (Callithrix jacchus)

    Neurobiology of Aging 74, 1-14 · 2019

    Google Scholar

  • Article

    The detrimental influence of attention on time-to-contact perception

    Attention, Perception, & Psychophysics 80 (6), 1591-1598 · 2018

    Google Scholar

  • Article

    Parkinson’s patients can rely on perspective cues to perceive 3D space

    Brain Research 1663, 161-165 · 2017

    Google Scholar

  • Conference abstract

    V6 is active during antero-posterior but not in lateral galvanic vestibular stimulation

    Journal of Vision 15 (12), 859-859 · 2015

    Google Scholar

  • Article

    Peripheral optic-flow involvement in the multisensory construction of the 3D percept in humans

    Frontiers · 2013

    Google Scholar

  • Journal abstract

    Alteration of three-dimensional vision in Parkinson’s disease: 551

    Movement Disorders 28, S195 · 2013

    Google Scholar

  • Journal abstract

    Altération de la vision tridimensionnelle dans la maladie de Parkinson

    Revue Neurologique 169, A127-A128 · 2013

    Google Scholar

  • Article

    Effect of audiovisual training on monaural spatial hearing in horizontal plane

    PLOS ONE 6 (3), e18344 · 2011

    DOI Google Scholar

  • Conference abstract

    The role of vergence in the multisensory construction of the 3-D percept in humans

    Perception 40 (1), 122-122 · 2011

    Google Scholar

Demos

Interactive examples

Technical demos aligned with data work (complexity, structure) and modeling — useful when scoping an algorithmic project or a scientific check.

Quick Sort visualizer

Classic sorting behavior — helpful to reason about CPU / memory cost at scale (ETL, features, sports data or sensors).

Quick Sort in short:

  1. Pick a pivot (here, the last element)
  2. Partition: move every element smaller than the pivot to the left, larger to the right
  3. Recurse on the left and right sub-arrays

Complexity: O(n log n) average, O(n²) worst case

Ready — click “Generate array” to begin
Pivot
Comparing
Sorted
Pending
Comparisons 0
Swaps 0
Elapsed 0ms
Elements 0

3D visualization

WebGL / Three.js — structures and networks (bio-inspired, graphs): useful to explain relational data or spatial signals.

α-helix (schematic)

Regular spiral of residues — a pedagogical model close to an α-helix. Rotate and zoom to explore.

Contact

Data project, sports / health app, or research?

Describe your context (available data, users, regulatory constraints), the business goal, and timing — I usually reply within a few business days.

Availability
Data POCs, field apps, pipeline reviews, and longer engagements