Researcher · Builder · Science communicator

Thomas
de Chillaz

I build ways to see complex systems more clearly — from the inner life of a cell to the surface of another world.

Thomas de Chillaz

Currently AI & Data student
Researching at MIT CSAIL

Use arrow keys to exploreSwipe to explore

01 / Health & computational biology

Finding signal
inside the cell.

At MIT CSAIL's Computational Biology Group, I am working where AI, interfaces, and genomics meet — making complex scientific knowledge easier to explore.

MIT Computational Biology Group
Single-cellData visualization

Single-cell RNA sequencing

Building web-based ways to navigate high-dimensional single-cell data, turning dense expression matrices into more legible biological patterns.

LISN LaboratoryGif-sur-Yvette

Multimodal survival analysis

Developing survival prediction from censored clinical data, then evaluating how CT scans, histopathology, and electronic health records can improve practical risk estimation.

CTPathologyEHRRisk

02 / Astronomy

Teaching machines
to read worlds.

Astronomy is my longest-running curiosity: a place to combine physics, data, and imagination with questions that stretch beyond one discipline.

EXP — 01

Exoplanet Habitability Classifier

Machine learning on NASA's Kepler dataset to estimate whether planetary conditions could be compatible with life.

  • Kepler
  • Feature engineering
  • Classification

LAND — 02

Landing Suitability Algorithm

A computer-vision system that combines terrain segmentation and crater detection to score safer planetary landing zones.

  • CNN
  • Segmentation
  • Hazard detection

“What is worth doing, is worth doing well.”

— My grandfather

03 / Education

Learning across
the boundaries.

2025 — 2029

01

CentraleSupélec × ESSEC

Bachelor of AI, Data & Management Sciences

An interdisciplinary program connecting rigorous technical work with strategy, organizations, and real-world decision-making.

  • Linear algebra
  • Calculus
  • Statistics
  • Optimization
  • Data structures
  • Data analysis

2023 — 2025

02

École Jeannine Manuel

French International Baccalaureate

Graduated with highest honors, specializing in mathematics, physics, and chemistry with expert-level mathematics.

2024

03

Stanford Online

Understanding Einstein: Special Relativity

04 / Selected work

Research that
becomes a product.

I learn fastest by building — models, agents, and tools that make a difficult process more understandable or useful.

01Medical AI

Lung Cancer Subtype Classifier

A CT-scan classifier that improved from 85.56% to 93.02% accuracy.

Open project notes

I refined EfficientNet-based architectures and redesigned augmentation so clinically relevant tumor regions survived the training pipeline.

  • EfficientNet
  • PyTorch
  • CT imaging
02Clinical workflow

Clinical Trial AI Agent

An automated call-center concept for faster clinical-trial patient classification.

Open project notes

The system combines conversational voice AI, internal clinical data, and lab models to structure a complex screening workflow for clinicians.

  • AI agents
  • ElevenLabs
  • Clinical data
03EdTech

Next Kareer

An AI preparation platform for stronger, more targeted job applications.

Open project notes

Developed in the CentraleSupélec accelerator loop, the product brings together interview practice, résumé feedback, and application tracking around the candidate journey.

  • Product
  • AI interviewer
  • Accelerator
04Voice AI

Intervue AI

A real-time mock-interview prototype with voice interaction and résumé context.

Open project notes

Built as a working demonstration with speech-to-text, text-to-speech, résumé parsing, and rapid response generation.

View public project
  • Flask
  • Gemini
  • ElevenLabs
60M+

05 / Science communication

Curiosity,
at scale.

I founded The Curious Minds to turn mathematics and astrophysics into stories people want to share. Across TikTok, Instagram, and YouTube, that work has reached 60M+ views.

03

Major social platforms

02

Languages spoken natively

01

Companion science website

Toolkit

How I work.

Research computing

  • Python
  • PyTorch
  • Torchvision
  • scikit-learn
  • NumPy
  • Pandas

Methods

  • Computer vision
  • Survival analysis
  • Multimodal ML
  • Data visualization

Product

  • HTML / CSS / JavaScript
  • SQLite
  • AI agents
  • User research

Open to ambitious ideas

Let's build what
doesn't exist yet.