Jean-Maximilien Cadic

Our patented technology is trusted by top tech companies in Japan, from digital business to telecom. Kironics is an industry leader in Generative AI. By using a fully data-driven appproach, Kironics provides advances solution in AI to support your business development.

• Distributed p2p framework for cloud computing • QUIC protocol over UDP for fast message passing • Network resources allocation powered by genetic algorithms • Trading system for computing resources

Developed buyer/seller segmentation model to improve personalization and reranking in SRE. Developed PII model in Japanese to ensure customer privacy. Optimize Mercari Search data pipelines with BigQuery, Delta lakes, and testing to reduce query latency. Improve the quality, security and performance of the product with good practices (CI/CD, Docker, DataDog). Define KPI and SLAs using stress tests, benchmark, and monitoring tools. Enforce design patterns in MLOps and promote good practices in the teams.

Lead the development and implementation of recommender systems and campaign forecasting for Catalina Japan (AEON, 7&11, COCOKARA...) using Neural Matrix Factorization, Transformers architectures. CI/CD using Azure DevOps, deployment of productions model using MLFlow on AzureML. Design and implement Databricks pipeline for Field Analytics team. Implement optimal graph based approach for customer identification (graph theory, optimization). Build a price forecast solution for short to long term prediction (Kalman filter, ARIMA, Prophet).

• Develop innovative solutions for client work built into standardized methods. • Apply machine learning and AI methods into client work spanning a range of use cases including exploratory insights, predictive modeling reporting, and optimization. • Development of patented solution for the generation of dental mesh. • Supervision of 15 engineers (project management, axis of research, code review...). • DevOps/MLOps, system engineer (Linux/Docker/Network), data analyst and model architect. • 3D Mesh generation by DNN, system developer, data analyst and model architect. • Earthquake prediction based on EMD/IMF and LSTM neural network. • Brain anomaly detection based on KL divergence using VAE. • Building reconstruction based on genetic algorithms. • Setup of server Linux and services AWS/GCP/AZURE.