Freelancer profile translated to English.

Description

I help startups and SMEs build and deploy AI applications and web solutions in production — LLM agents, RAG pipelines, chatbots, API integrations (Gemini, Mistral, Claude), and custom websites.

What I offer:

Architecture and development of AI agents / RAG systems

Integration of LLMs into web applications (Next.js, React, FastAPI)

Full-stack web development: showcase websites, business applications, dashboards

Cloud deployment AWS · Docker · Nginx — production-ready

API cost audit and optimization (latency vs. quality)

How I work: short or long assignments, full-remote, clear communication, documented deliverables.

📩 Contact me to discuss your project.

Languages

French
Native or bilingual
English
Fluent

Workplace preferences

Remote only

Primarily works remotely

Ubi Solutions
Senior AI Engineer
October 2025 - March 2026 (5 months)
Paris, France
Design and development of a multi-agent AI backend for a SaaS chatbot and data prediction application.
Development of a multi-agent system with LangGraph ReAct: state graphs, tool execution, and human-in-the-loop supervision.
Dual API architecture with FastAPI and SSE streaming, enabling independent scaling between the chatbot and the prediction engine.
Development of advanced LLM tools: text-to-SQL (Gemini / Snowflake Cortex), anomaly detection, semantic glossary, document search.
XGBoost prediction system with autoregressive forecasting, confidence intervals, and alert thresholds.
Snowflake integration with private key authentication, multi-client semantic layers, and complete data isolation.
Delivery of a Next.js frontend, PostgreSQL / Prisma persistence layer, and public OpenAPI contract.
Deployment on AWS EC2 with Docker Compose, elastic IP management, and dev / staging / production environments.
Docker Next.js FastAPI Amazon Web Services Machine Learning
F2J Industry
AI Engineer
April 2024 - September 2025 (1 year and 5 months)
Mission 1 – Optimization of industrial processes through AI

Development of a conversational AI agent (chatbot and voicebot) to assist operators in real-time production monitoring.
Centralization of production data (OEE, availability, quality) in an SQL database.
Implementation of a FAISS index with ANN search for contextualization.
Deployment of a RAG system for real-time KPI analysis → 15% reduction in machine downtime and 5% increase in OEE.

Mission 2 – Development of AI-based industrial vision systems

Fine-tuning and deployment of YOLOv8 for automated quality control → 20% improvement in weld defect detection.
Construction of a production label recognition pipeline with EasyOCR and development of an operator interface (PyQt) integrated into the production flow.
Fine-tuning of Depth-Pro and Depth-Anything V2 with normalization optimization → Research Tax Credit valuation: €112,000.
Improvement of the Shape-from-Shading method for 3D reconstruction → R&D Tax Credit valuation: €97,000.
Development of a Random Forest model for estimating the number of pallets from depth data.
Python FastAPI PostgreSQL Random Forest Computer Vision
Upwork
Data Scientist
September 2022 - April 2024 (1 year and 7 months)
Development of a user assistant chatbot based on Natural Language Processing (NLP) and Deep Learning.

🔗 GitHub: github.com/Alexmestrah/ChatBot_By_Alex
LLM Python NLP Deep Learning