Corentin Juste.

Solution Engineer · Data & AI

I graduated from ESILV as an engineer specialising in Data & AI, and I’m now in the final year of the IASD Master’s (AI, Systems, Data) at Université Paris Dauphine-PSL.

Alongside my studies, I work as a Solution Engineer at ChapsVision on ArgonOS, a large-scale data processing platform used by public and private organisations. I deploy it at client sites, build ETL and AI components, and provide technical support throughout client projects.

Before that, I spent two years at CCF as an apprentice Data Scientist, measuring the causal impact of CRM campaigns and classifying banking transactions with NLP.

CVContact
CJ
Corentin Juste
01 / Experience

Career

Click on an experience to see the details.

02 / Blog

Articles

Deep Learning Specialization: what 5 Andrew Ng's courses taught me

How Andrew Ng's Deep Learning Specialization bridged the gap between academic theory and the practice of deep learning.

Deep LearningRéseaux de neuronesOptimisationCNNRNN
Read article →

Teaching an RL agent to play 2048: from the 512 Tile to 4096

A hands-on account of training a reinforcement learning agent: the dead ends, the right tools, and the lessons you can only learn by doing.

PythonReinforcement LearningPyTorchStable-Baselines3GymnasiumTensorBoard
Read article →

Building an AI Help Center in 3 Days: RAG and Leadership at a Hackathon

How our team designed a RAG pipeline to transform a university help center, and what I learned leading four developers under time pressure.

HackathonRAG (Retrieval Augmented Generation)ElasticsearchEmbeddingsNLPFastAPILeadershipProject managmentPython
Read article →
03 / Portfolio

My projects

TerraBot - Infrastructure as Conversation
LangGraphTerraformGraphQLLLMPythonStreamlit

TerraBot - Infrastructure as Conversation

Chatbot that turns cloud infrastructure requirements into Terraform with real-time cost estimates.

GitHub
2048 - Reinforcement Learning
PythonReinforcement LearningPyTorchStable-Baselines3GymnasiumTensorBoard

2048 - Reinforcement Learning

Deep reinforcement learning agent playing the 2048 game using MaskablePPO, a custom CNN and reward shaping

GitHub
SatGen - Satellite Image Generation
PythonPyTorchDeep LearningGANComputer VisionHugging FaceTensorBoard

SatGen - Satellite Image Generation

Realistic satellite image generation from road and building segmentation masks

GitHub
Voice Assistant AI Agent
PythonLangChainLangGraphRAG (Retrieval Augmented Generation)PostgreSQLChromaDBTwilio

Voice Assistant AI Agent

Multi-agent voice assistant for the restaurant industry, with multilingual support, phone integration, RAG and analytics dashboard

GitHub
IBM Hackathon Winner - RAG Help Center Chatbot
RAG (Retrieval Augmented Generation)ElasticSearchEmbeddingsNLPFastAPIMariaDBPython

IBM Hackathon Winner - RAG Help Center Chatbot

Intelligent BI pipeline developed during an IBM hackathon: semantic search, RAG and analytical dashboard

GitHub
Customer Revenue Prediction
PythonScikit-learnLightGBMXGBoostTensorFlowFeature Engineering

Customer Revenue Prediction

Per-session revenue prediction for the Google Merchandise Store using Google Analytics data

GitHub