Freelancer profile translated to English.
Back to original languageAbout Yanis
🎯 Senior AI/ML Engineer (PhD ENS Paris) — GenAI, MLOps
I help companies design and deploy advanced, robust, and scalable AI systems, with a focus on business value, GDPR compliance, and operational excellence.
🧠 GenAI: Secure autonomous agents, LangGraph, RAG, MCP, google adk, langchain, evaluation, security
⚙️ Cloud/MLOps: GCP (Vertex AI, GKE, Cloud Run), CI/CD, Docker, FastAPI, monitoring
🤖 80% automation via GenAI agents
📦 Open source contributor (NeuralForecast – Nixtla)
📍 Available for freelance (France / Remote) for strategic AI, GenAI, or forecast missions.
Test my AI at
French
Native or bilingual
English
Fluent
Can work on-site
Paris (up to 20km)
Experience
- DecathlonSenior AI Engineer - GenAI & Agentic SystemsSPORTSFebruary 2025 - Today (1 year and 4 months)Paris, France
Autonomous AI Agents (GenAI) & Automation
1. Agent Onboarding Wholesale (End-to-End)Context:Manual and very slow B2B customer onboarding process (3 weeks). Slowed down time-to-revenue.Solution:Intelligent and multimodal chatbot integrated with Google Chat to guide/automate new customer onboarding.- RAG Assistant: Answers questions about the customer onboarding process.
- Automatic Onboarding: Extracts and analyzes contracts/documents (PDF, Images) to automatically populate data after validation.
- Orchestration: Google Chat, Vertex AI, BigQuery, secure proxy (VPC) on GCP, SAP Write.
Impact:Drastic acceleration (3 weeks -> a few minutes) and complete autonomy.2. IT Support Agent (LangChain/LangGraph)**Context**: Support overwhelmed by repetitive "price mismatch" tickets.Solution:Complex pipeline with LangChain/LangGraph agent orchestrated by Airflow on AWS. Document analysis and automatic extraction.Impact:80% of tickets resolved automatically, teams freed up.Stack:Python, Vertex AI, Google ADK, Cloud Functions, LangGraph, BigQuery, Airflow, RAG, NeuralForecast, Pytorch - DECATHLONSenior Data Scientist - Forecasting & Supply Chain OptimizationSPORTSJune 2024 - February 2025 (8 months)Paris, FranceSales Forecasting & Open-Source ContributionsContext & Challenge:Optimize European inventory through more accurate forecasts. Traditional models lacked explainability for business users.Technical Solution (Framework Expertise):In-depth modification of the TFT architecture (NeuralForecast/Nixtla source code).Development of native interpretability modules (Feature Importance, Attention Weights) to make AI auditable.Open-Source Contributions (Integrated into the official library):PR #1230 Merged (Native Interpretability) :PR #1104 Merged (Architecture Improvements) :Business Impact:Accuracy: -2.5% error (WAPE).ROI: ~ Several M€ in annual savings through inventory optimization.
- FeedgySenior Data Scientist | MLOpsENERGY AND UTILITIESNovember 2023 - April 2024 (6 months)Paris, FrancePhotovoltaic Power Plant Production Forecasting (End-to-End)Context & Challenge: Industrialize the prediction of photovoltaic power plant production to optimize energy management. The main challenge was to structure a rich R&D approach into a robust MVP product capable of handling complex time series (weather and energy data) and scaling on the cloud.Solution:
- Architecture & MLOps: Design and deployment of a complete machine learning pipeline and orchestration on Airflow. Implementation of monitoring for data drift and concept drift.
- Modeling: Training and optimization of boosting models (XGBoost, LightGBM) and Deep Learning on time series.
- Management: Roadmap steering (Jira), MVP definition, and technical supervision of a junior Data Scientist team.
Impact:- Successful production deployment of the automated forecasting pipeline. Reliable delivery through model quality monitoring (Drift detection).
- Team upskilling on MLOps standards.
Technical Environment: Python, AWS (SageMaker, EKS), Kubernetes, ArgoCD, Airflow, MLflow, Evidently, Docker, XGBoost/LightGBM, GitHub Actions.
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Master 2 Artificial Intelligence, Systems & DataUniversité Paris Dauphine - PSL2021M2 informatique: formation complémentaire centrée sur l'industrialisation de l'IA
- PhD in MathematicsÉcole normale supérieure2019
Certifications
- Deep Learning SpecializationCoursera2020