AI Engineer · PhD Candidate · INSA-Lyon

Aymar
Tchagoue

Designing intelligent systems at the intersection of machine learning and materials science. Building the future of AI-assisted polymer discovery at LIRIS/IMP, INSA-Lyon.

Aymar Tchagoue presenting at Starthèse
3+
Publications
3+
Years Research
4+
Courses Taught
5+
Engineering Experience
Who I Am

AI researcher

I am a PhD candidate at INSA-Lyon, working within the LIRIS and IMP laboratories. My research focuses on AI-driven polymer design, combining knowledge modeling, counterexample analysis, property prediction, and guided molecular generation to accelerate the discovery of new materials.

Before my PhD, I explored a diverse mix of AI, design, and tech. At GEOVECTORIX, I conceived a virtual and augmented reality platform for museums, combining spatial analysis with immersive design. Then, as a design engineer at TheDIGITALITE, I worked on parametric designs across various architectural projects. Earlier, at LIRIS, I explored federated learning for predictive maintenance, training models to predict the remaining useful life of turbojet engines in a privacy-preserving setup. I hold a background in AI, computer vision, and robotics engineering.

Alongside my research, I teach courses on data science, human-computer interaction, and the digital society at CPE-Lyon and INSA-Lyon. I’m passionate about making cutting-edge AI both accessible and impactful.

Lyon, France
PhD Defense: June 2026 · ED InfoMaths 512
Career

Work Experience

A journey from engineering in Cameroon to cutting-edge AI research in Lyon.

Jan 2024 – Present
PhD Candidate
LIRIS/IMP · INSA-Lyon · ED InfoMaths 512 · Lyon, France
Thesis: "AI for Polymer Design : From Knowledge Modeling–Based Counterexample Analysis to Guided Molecular Generation". My research focuses on AI-driven polymer design, combining knowledge modeling, counterexample analysis, property prediction, and guided molecular generation to accelerate the discovery of new materials.
Feb 2023 – Present
Teacher
CPE-Lyon & INSA-Lyon · Lyon, France
  • Data Mining & Big Data — 3ETI, IRC & ICS · CPE-Lyon (2023)
  • Human-Computer Interaction — 4IFA · INSA-Lyon (2024–2025)
  • Computer Science & Digital Society — ISN, FIMI 1–2 · INSA-Lyon (2026)
Nov 2022 – Dec 2023
Research Engineer
CNRS · LIRIS UMR 5205 / IMP UMR 5223 · Lyon, France
Led the construction of the first AI-driven literature pipeline for polymer data extraction. Designed and built the Epoxy-Amine database from thousands of scientific papers, enabling downstream machine learning models for property prediction.
Feb 2022 – Aug 2022
Research Engineer Intern
INSA-Lyon / LIRIS · Lyon, France
Explored federated learning for predictive maintenance. Built models for remaining useful life prediction of turbojet engines using the NASA CMAPSS dataset, achieving competitive accuracy in a privacy-preserving federated setup.
Sep 2021 – Jan 2022
Design Engineer
TheDIGITALITE · Yaoundé, Cameroon
Delivered parametric design and domotics projects. Combined CAD modeling with smart home system integration for residential and commercial clients.
Feb 2021 – Jul 2021
Developer Engineer Intern
GEOVECTORIX · Douala, Cameroon
Designed and built a virtual and augmented reality platform for the Cameroon National Museum. Conducted cartographic analysis of Yaoundé using satellite and drone imagery processed with ArcGIS.
Research Output

Publications

All three papers are accepted and published — a milestone I'm genuinely proud of.

✓ Accepted · 2025
JCIM — Journal of Chemical Information and Modeling
Dual Embedding: A Fine-Tuned Language Model Approach for Accurate Polymer Glass Transition Temperature Prediction
Aymar Tchagoue et al. · 2025
A fine-tuning approach combining dual molecular embeddings with BERT-based LMs to achieve state-of-the-art Tg prediction across diverse polymer families.
✓ Accepted · 2026
DKE — Data & Knowledge Engineering
CCASL: Counterexamples to Comparative Analysis of Scientific Literature — Application to Polymers
Aymar Tchagoue et al. · 2026
A novel framework using counterexample-driven logic to support automated comparative analysis of scientific claims, applied to polymer property literature.
✓ Accepted · 2026
JCIM — Journal of Chemical Information and Modeling
Multi-Score Reinforcement Learning for High-Tg Polyimide Design
Aymar Tchagoue et al. · 2026
A reinforcement learning strategy leveraging multiple reward signals to guide the de novo molecular generation of high glass-transition-temperature polyimides.
Paper
DOI
Toolkit

Skills & Languages

Programming
Python C/C++ MATLAB SQL PHP Linux Pandas NumPy KNIME ...
Artificial Intelligence
Deep Learning Machine Learning Graph Neural Networks Transformers Representation Learning Reinforcement Learning LLMs / Fine-tuning TensorFlow PyTorch Flower Knowledge Modeling ...
AI for Materials & Polymers
Polymer Informatics Molecular Descriptors SMILES PSMILES BigSMILES Property Prediction Modeling De novo Molecular Generation Experimental Design Statistical Modeling RDKit ...
Engineering & Design
AutoCAD SolidWorks Revit Rhinoceros Proteus Raspberry Pi ...
3D & Geospatial
ArcGIS Unity Unreal Engine AR/VR Oculus VR ...
Languages
🇫🇷 Français
Full Professional Proficiency
🇬🇧 English
Full Professional Proficiency
Research Communication
IMP High-Throughput Characterization Platform Inauguration · Oct 2025
Oral presentation — AI and Polymers: The Role of High-Throughput Strategies in Accelerating Materials Discovery
IMP-Days · Apr 2024
Poster — Toward AI-Based Polymer Discovery: Epoxy/Amine Data Collection from Scientific Literature
GFP Conference · Nov 2023
Oral presentation — Towards AI-Based Polymer Discovery: an Optimized Data Collection System from Epoxy/Amine Tables
Colloque Chimie & IA · Feb 2023
Participation
Academic Background

Education

AI, Vision and Robotics Engineer
National Higher Polytechnic School of Douala · Cameroon
Sep 2016 – Aug 2021
Five-year engineering degree covering artificial intelligence, computer vision, robotics, embedded systems, and control theory. Graduated with a strong foundation in applied AI and engineering design.
PhD in AI
INSA-Lyon · ED InfoMaths 512 · LIRIS
Jan 2024 – Jun 2026 (expected)
Doctoral research at the intersection of knowledge representation, machine learning, and polymer chemistry. The thesis introduces novel counterexample analysis and RL-guided generation methods for material design.
Academic Teaching

Courses I Teach

Sharing knowledge in data science, HCI, and digital society with engineering students.

Data Mining & Big Data
3ETI, IRC & ICS · CPE-Lyon · 2023
Human-Computer Interaction
4IFA · INSA-Lyon · 2024–2025
Computer Science & Digital Society
ISN, FIMI 1–2 · INSA-Lyon · 2026
Let's Connect

Get in Touch

If you’d like to discuss, I’d love to hear from you.