Freelancer profile translated to English.
Back to original languageAbout Saliou
Most AI consultants make recommendations.
I build AI systems that work in real-world conditions — and deliver measurable results.
🏥 Hospital Sector: Medical AI & LLM
Design and deployment of an intelligent patient data analysis system based on LLM, RAG, and LangChain, capable of processing complex medical data (EHR, lab results, imaging).
📊 Results:
+65% accuracy in identifying medical patterns vs. traditional analysis
-40% patient record analysis time
+50% relevance of therapeutic recommendations
x3 processing speed of heterogeneous medical data
100% operational AI pipeline ready for production
What makes the difference:
I don't propose "theoretical" architectures.
I design, develop, and run them myself on complex cases (health, critical data, multi-source).
🛠️ Expertise:
Generative AI • LLM • RAG • NLP • Computer Vision • Agentic Systems • Data/AI Pipelines • Production Deployment
💼 Engagement Models:
🏗️ Build: POC → MVP → Production
🔍 AI Audit: Analysis + Roadmap + Prototype in 2 weeks
⟳ Support: Rapid iterations, ROI focus
⚙️ Domains & Technologies
LLM • RAG • GPT • Claude • Mistral • Llama • DeepSeek • Qwen • Gemini • Gemma • Phi • Kimi • Grok
Fine-tuning • LoRA • vLLM • CUDA • Transformers
LangChain • LlamaIndex • CrewAI • AutoGen • LangGraph
Python • FastAPI • Docker • Kubernetes • MLOps • LLMOps • DevOps
Embedding • Vector DB • Faiss • Chroma • Milvus • Qdrant • Pinecone
Computer Vision • CNN • OpenCV • YOLO
NLP • Text Mining
Stable Diffusion • GenAI • Text-to-Speech • Whisper
AI Agents • MCP
AWS • Azure • GCP
Machine Learning • Deep Learning • PyTorch • TensorFlow • HuggingFace
PostgreSQL • PySpark • Airflow • Kafka
Advanced Statistics • Time Series • Modeling
French
Native or bilingual
English
Conversational
Can work on-site
Toulon (up to 50km), Marseille (up to 50km), Toulouse (up to 50km), Lyon (up to 50km), Paris (up to 50km)
Experience
- Nando ITHealthcare Mission - Patient Data Analysis with LangChainMEDICALOctober 2025 - December 2025 (2 months)Toulon, FranceAnalysis of patient data in a hospital setting using LangChain to improve diagnosis and treatment planning. Processing massive volumes of medical records, identifying patterns and correlations that might be missed by human analysis. Early disease detection and development of personalized treatment strategies.LangChain Pipeline: Medical data processing architecture based on LangChain for dynamic content workflows and intelligent insightsEarly Detection: Pattern and correlation analysis in medical records for early disease identificationTherapeutic Recommendations: Generation of personalized treatment recommendations through LLM-augmented content generationDemo Dataset: Functional prototype on a fictional dataset simulating real hospital records (diagnoses, treatments, patient histories)Multi-Source Integration: Connection of heterogeneous data sources (EHR records, lab results, medical imaging) via LangChain connectors
- Nando ITMultimodal RAG Project Mission - DeepSeek Janus-Pro & VisionE-COMMERCEJanuary 2026 - March 2026 (2 months)Toulon, FranceImplementation of a multimodal RAG system leveraging DeepSeek's latest model, Janus-Pro, for intelligent querying of documents combining text and images. Complete architecture integrating visual encoding, vector database, and interactive web interface.Multimodal LLM: Integration of DeepSeek Janus-Pro as a multimodal language model for understanding and generation combining text and visual contentVisual Encoding: Use of ColPali as a vision encoder for indexing and semantic search of image-rich documentsVector Database: High-performance storage and retrieval of multimodal embeddings via Qdrant, with vector similarity searchRAG Pipeline: Complete pipeline for document ingestion, multimodal chunking, embedding, indexing, and retrieval-augmented generationWeb Interface: Interactive application deployed with Streamlit for intuitive querying of the multimodal knowledge base
- Nando ITCustom MCP Server Mission - In-depth Web Search & RAGRESEARCHJune 2025 - August 2025 (2 months)Toulon, FranceDevelopment of a custom MCP (Model Context Protocol) server, connected to Cursor, enabling in-depth web searches via Linkup and integrating a RAG system based on LlamaIndex for retrieval and data-augmented content generation.Custom MCP Server: Design and development of a tailor-made MCP server exposing AI tools accessible directly from the Cursor environmentIn-depth Web Search: Integration of the Linkup API for advanced semantic web searches, retrieval of reliable sources, and automatic content summarizationRAG Pipeline with LlamaIndex: Complete RAG architecture (indexing, chunking, embedding, retrieval) via LlamaIndex for querying custom knowledge basesCursor Integration: Native connection to Cursor's MCP protocol, allowing developers to access search and RAG capabilities directly within their IDEAI Orchestration: Combination of real-time web search + RAG for enriched, factual, and contextualized answers across any domain
Recommendations
Be the first to recommend Saliou
Help this freelancer shine by sharing your experience working together.
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
- DevOps EngineerDevUniversity - Certified Paris 1 Panthéon Sorbonne University2024450 h de Formation Administrateur Système DevOps. Développement d'application avec Python : Fast API, Linux et Bash, Virtualisation et Vagrant. Systèmes Linux et Agilité : Administration des systèmes Linux, Nginx, Structure de données et Stockage. Base de données : SQL, PostgreSQL, NEO4J, Elasticsearch, MongoDB. CI/CD : Git, Github, Jenkins, Docker, Gitlab, Kubernetes. Cloud AWS : Praticien du Cloud AWS, Architecte de Solutions AWS. Automatisation et monitoring : Terraform, Ansible, Prometheus, Grafana. Cloud DevOps : AWS-Lambda, AWS-API Gateway, AWS-CodeFormation, AWS-CodePipeline, Migration Vers le Cloud, etc.
- Data Scientist, co-certified diploma by MINES ParisTech PSLDataScientest.com2023Analyse et visualisation des données : Numpy, Pandas, Matplotlib, Seaborn, Power BI Techniques de Machine Learning : Tensorflow – Keras, Machine Learning avec PySpark, Modèles et algorithmes de classification, Méthodes de clustering, Méthodes de régression, Méthode de réduction de dimension, Séries temporelles, NLP, Deep Learning, Computer Vision
Certifications
- Develop an Artificial Intelligence Solution (Machine and Deep Learning)Ascencia Business School - DataSCientest2023
- Data ScientistDataScientest.com2023