08/05/2026 updated
100 % available
AI & Machine Learning Engineer | RAG Systems | LLM & Full-Stack Development
Grenoble, France
Worldwide
Master of Science in Informatics and Master of Artificial Intelligence (Grenoble INP - ENSIMAG)About me
AI Engineer specializing in Generative AI, RAG, and agentic AI systems. I build production-ready AI applications with Python, FastAPI, LangChain, and LangGraph, focusing on reliable automation, semantic search, and scalable backend solutions.
AI AgentsRAGLLMPython (Programming Language)Machine LearningNext.jsReact.jsGitBackend ProgrammingDocker
AI & LLM Frameworks
Expertise in PyTorch, Hugging Face (Transformers, PEFT/LoRA), LangChain, LangGraph, LlamaIndex, vLLM, and RAGAS for building and fine-tuning large language models and AI pipelines.
Retrieval-Augmented Generation (RAG) Systems
Design and deployment of enterprise RAG systems unifying heterogeneous knowledge bases, building hybrid retrieval engines combining dense semantic vectors and sparse lexical search, and establishing multi-level RAG evaluation frameworks.
Web & Backend Development
Development of full-stack applications and RESTful APIs using FastAPI, Flask, Node.js, Express.js, Next.js, React, and REST APIs, including secure authentication, authorization, and CRUD operations.
Machine Learning & Data Processing
Proficiency in scikit-learn, NumPy, Pandas, Matplotlib, and BGE-M3 (Hybrid Embeddings) for data analysis, feature engineering, and machine learning workflows.
Databases & Vector Stores
Experience with Qdrant, FAISS, ChromaDB, PostgreSQL, Neo4j, and MongoDB for structured and unstructured data storage and vector-based retrieval.
Cloud, DevOps & Data Pipelines
Hands-on experience with AWS (Lambda, S3, EC2), Docker, Nginx, Git, and Prefect for ETL orchestration, automated data pipelines, and cloud infrastructure management.
Programming Languages
Strong command of Python, JavaScript/TypeScript, and SQL for developing AI solutions, backend services, and data processing scripts.
Intrusion Detection & Security Systems
Analysis of Host-based and Network-based Intrusion Detection Systems, ensemble fusion methods, and evaluation on datasets such as ADFA-LD and CIC-IDS-2017.
Expertise in PyTorch, Hugging Face (Transformers, PEFT/LoRA), LangChain, LangGraph, LlamaIndex, vLLM, and RAGAS for building and fine-tuning large language models and AI pipelines.
Retrieval-Augmented Generation (RAG) Systems
Design and deployment of enterprise RAG systems unifying heterogeneous knowledge bases, building hybrid retrieval engines combining dense semantic vectors and sparse lexical search, and establishing multi-level RAG evaluation frameworks.
Web & Backend Development
Development of full-stack applications and RESTful APIs using FastAPI, Flask, Node.js, Express.js, Next.js, React, and REST APIs, including secure authentication, authorization, and CRUD operations.
Machine Learning & Data Processing
Proficiency in scikit-learn, NumPy, Pandas, Matplotlib, and BGE-M3 (Hybrid Embeddings) for data analysis, feature engineering, and machine learning workflows.
Databases & Vector Stores
Experience with Qdrant, FAISS, ChromaDB, PostgreSQL, Neo4j, and MongoDB for structured and unstructured data storage and vector-based retrieval.
Cloud, DevOps & Data Pipelines
Hands-on experience with AWS (Lambda, S3, EC2), Docker, Nginx, Git, and Prefect for ETL orchestration, automated data pipelines, and cloud infrastructure management.
Programming Languages
Strong command of Python, JavaScript/TypeScript, and SQL for developing AI solutions, backend services, and data processing scripts.
Intrusion Detection & Security Systems
Analysis of Host-based and Network-based Intrusion Detection Systems, ensemble fusion methods, and evaluation on datasets such as ADFA-LD and CIC-IDS-2017.
Experience
Professional experience0 - 5 years
Freelancer since2026
Languages
EnglishFluentFrenchBasic knowledge
Project history
Architected and deployed an enterprise RAG system unifying 5 heterogeneous knowledge bases into a citation-backed Q&A assistant. Built a Hybrid Retrieval Engine using Qdrant achieving 83.3% Hit@5 retrieval rate. Engineered REST APIs using FastAPI and configured a local vLLM inference engine. Developed automated ETL data pipelines using Prefect and established a multi-level RAG evaluation framework.
Applied PEFT to fine-tune Gemma 2 for fast training on resource-limited hardware while preserving performance. Technologies used: PyTorch, Pandas, Hugging Face, Transformer.
Designed a reflection-pattern agent workflow to verify answers and perform re-retrieval for improved accuracy and reliability. Implemented hybrid retrieval combining BM25 and vector embeddings. Technologies used: LangGraph, LangChain, OpenAI, ChromaDB.