Hi,
I'm abdessamad đź‘‹
AI/ML and software engineer with a love for clean code and creative solutions.

Looking for a junior AI engineer role worldwide. Let's build something intelligent together.

AI

About

My name is Abdessamad Idboussadel, and I am currently in the final year of my Master's in Artificial Intelligence at Université Claude Bernard Lyon 1, specializing in AI and Machine Learning. I am also a graduate in Computer Engineering from ENSA Agadir, giving me a hybrid profile that combines strong expertise in AI with solid skills in software engineering, system architecture, and the development of complex applications. This unique combination allows me to approach intelligent agent projects in a robust, practical, and production-oriented way. I am the creator of Capptivo and the engineer behind EcoAI.

I am currently seeking a Junior or Entry-Level AI Engineer position where I can contribute to building real-world AI products while continuing to grow my expertise. I am particularly interested in roles involving:

  • AI Engineering & Generative AI
  • Machine Learning & Deep Learning
  • LLMs, RAG, AI Agents, and Multi-Agent Orchestration (LangChain / LangGraph)
  • AI-powered Full-Stack Applications
  • MLOps & AI Infrastructure

Work Experience

E

Ecofinance Analytics

Apr 2026 - Present
AI Engineer Intern
Lead engineer behind EcoAI, Ecofinance's conversational assistant for Moroccan financial analysis, from data collection and indexing to the deployed, production-ready product delivered to end users. Designed a full-stack monorepo architecture (scrapers, API, web interface) and a multi-agent system built on LangGraph: specialized agents, RAG, dedicated tools, skill routing, and generative UI in the chat. Optimized the end-to-end pipeline (metadata filters, hybrid search, semantic cache, prompt caching, connection pooling, etc.) to reduce average response time from 3-4 min to ~30 s without sacrificing answer quality. Technologies used: Python, FastAPI, LangGraph, LangChain, Next.js, NestJS, TypeScript, RAG, Docker, Redis, PostgreSQL.
V

VOID

Feb 2025 - Jul 2025
Software Engineer Intern
Designed and developed a web app for Elsan, a leading private healthcare group in France, using a headless architecture with Drupal 10 as the CMS and Next.js 14 for the front end. The project aimed to handle high traffic (~8 million views/month) while improving performance and SEO. It was carried out following the Agile SCRUM methodology, fostering continuous collaboration and rapid adaptation to client needs. Technologies used: PHP, JavaScript, Drupal, Next.js 14, Tailwind CSS, MySQL, Docker, Git, Drush.
V

Vala Bleu

Feb 2024 - Jun 2024
Software Engineer Intern
Developed a comprehensive medical clinic management application over 5 months. The application simplifies appointment scheduling, medical procedures management, prescription handling, billing, and reimbursement processes. Technologies used: ReactJS, Laravel, JavaScript, Zustand, Tailwind CSS, PHP, MySQL.

Skills

Programming Languages

Python
JavaScript
PHP
HTML
CSS
TypeScript
Java
Node.js

Frameworks/Libraries

LangChain
LangGraph
PyTorch
TensorFlow
FastAPI
scikit-learn
NumPy
Flask
Pandas
React.js
Laravel
Next.js
Spring Boot
TailwindCSS
Hugging Face
OpenCV
Keras
NLP
Shadcn UI
Drupal
Optuna
MLflow
Pytorch Lightning

Databases

MySQL
PostgreSQL
MongoDB
Qdrant
Pinecone

Methodologies/Tools

Git
GitHub
BitBucket
Docker
Postman
REST API
MCP
GraphQL
SOAP
tRPC
Agile/SCRUM
Microservices
AI Projects

Check out my AI work

I've worked on a variety of AI projects, from machine learning models to deep learning applications. Here are a few of my favorites.

EcoAI

EcoAI

EcoAI is Ecofinance Analytics' production conversational assistant for Moroccan financial analysis. Users ask questions in natural language and get grounded answers backed by indexed market data, company filings, and financial sources. Built end to end as a full-stack monorepo with scrapers, a FastAPI API, and a Next.js chat interface, powered by a LangGraph multi-agent system: specialized agents, RAG, dedicated tools, skill routing, and generative UI components rendered inline in the conversation. Pipeline optimizations across metadata filters, hybrid search, semantic cache, prompt caching, and connection pooling cut average response time from 3-4 minutes to around 30 seconds without sacrificing answer quality.

Python
FastAPI
LangGraph
LangChain
Next.js
NestJS
TypeScript
RAG
Docker
Redis
PostgreSQL
ChatSimple

ChatSimple

SimpleChat is a full-stack platform that lets anyone build, customize, train, and deploy AI-powered chatbots on their website with a single script tag. The platform features AI chatbot creation with personalized instructions, document upload support and website crawling for knowledge bases, real-time chat capabilities, and a fully customizable embeddable widget. Advanced features include human handoff system, multi-user workspace collaboration with role-based access control, analytics dashboard, conversation management, and a usage-based credit system. Built with LangChain and LangGraph for AI orchestration, PostgreSQL for data storage, Pinecone for vector embeddings, and real-time WebSocket updates.

Python
Langchain
FastAPI
Next.js
Docker
TypeScript
TailwindCSS
Shadcn
Redis
Alembic
Celery
PostgreSQL
Pinecone
OpenAI
LangGraph
WebSockets
JWT
CodeForge

CodeForge

CodeForge is an AI-powered app builder in the spirit of Lovable, Bolt, and Cursor. Describe what you want in plain language, and the platform generates a complete, production-ready web application you can preview and refine in real time. Users iterate through conversation: tweak layouts, add features, fix bugs, and watch changes land instantly without leaving the browser. Refactored around a monorepo architecture with LangChain and LangGraph orchestrating a multi-agent pipeline. Specialized agents handle planning, code generation, file management, and validation, while an isolated E2B sandbox runs and previews the output safely. A Next.js frontend ties it all together with live preview, chat-driven editing, and a workflow that turns a rough idea into a deployable app in minutes.

Python
FastAPI
LangChain
LangGraph
Next.js
TypeScript
TailwindCSS
E2B Sandbox
DeepSeek API
Node.js
DeepAnimDance

DeepAnimDance

DeepAnimDance implements posture-guided person image synthesis, transferring motion from a source video to a target identity. The project explores multiple approaches: starting with a baseline Nearest Neighbor method, progressing through supervised learning with Vanilla Neural Networks (both vector-based and image-based), and culminating in GAN-based refinement using WGAN-GP with L1 loss for enhanced realism. The system uses MediaPipe Pose for skeleton extraction (33 landmarks reduced to 13 joints), and generates 64Ă—64 images in real-time. Includes a Flask web interface for browser-based visualization of the three-panel demo (source, skeleton, generated image).

Python
PyTorch
OpenCV
MediaPipe
Flask
NumPy
WGAN-GP
Computer Vision
Deep Learning
GANs
SentiTrade

SentiTrade

SentiTrade is a trading signals dashboard that combines multi-source sentiment analysis with Temporal Fusion Transformer (TFT) price prediction. Select a stock ticker to generate trading signals backed by aggregated sentiment and a 5-day forecast. The app includes model interpretability (attention weights, global and local feature importance) and empirical validation showing sentiment as a top-tier predictor.

Python
PyTorch
Flask
QLoRA fine-tuned Llama-2-7B
Qlora
TFT
Lightning
Hugging Face
Sentiment Analysis
JavaScript
TailwindCSS
Railway
Git
Management App for a Medical Clinic

Management App for a Medical Clinic

This project, developed in collaboration with Vala Bleu, involves creating a comprehensive management application for a medical clinic to streamline appointments, prescriptions, certificates, and billing processes. The application provides a complete digital solution that enhances operational efficiency, improves patient care coordination, and simplifies administrative tasks for healthcare professionals. I developed an intelligent assistance chatbot for the application, based on internal company data and scientific articles from the platform, using a RAG approach.

Python
FastAPI
Langchain
Pinecone
React
Laravel
PHP
JavaScript
TailwindCSS
Zustand
MySQL
Summize AI

Summize AI

It is an advanced full-stack AI application that simplifies the summarization process. With just a click, it summarizes any article with just a URL, PDF documents, Word files, and even PowerPoint presentations. This intelligent tool leverages cutting-edge AI technology to extract key information and provide concise, accurate summaries across multiple document formats.

JavaScript
React
TailwindCSS
Express.js
Material UI
Node.js
Redux
Full-Stack & Software Engineering

Full-Stack & Software Engineering Projects

I've worked on a variety of full-stack and software engineering projects, from simple websites to complex web/desktop applications. Here are a few of my favorites.

Capptivo

Capptivo

I built Capptivo, an open-source screen recorder for polished demos and a free alternative to Screen Studio and Cursorful. The cross-platform desktop app (macOS, Windows, Linux) features smart follow-cursor zoom, click-based auto zooms, on-screen annotations, custom cursors, and multi-format export, with native capture pipelines per OS, hardware H.264 encoding, and fully on-device editing. It supports 11 languages and has reached nearly 600 GitHub stars and over 6,000 downloads.

Rust
Tauri
TypeScript
React
TailwindCSS
FFmpeg
macOS
Windows
Linux
i18n
Optimize Stratégie

Optimize Stratégie

I developed a comprehensive multilingual website for Optimize Stratégie, a leading technical audit and energy consulting firm in Marrakech. The platform serves as both a corporate showcase and a lead generation tool, highlighting their expertise in technical audits, energy efficiency consulting, and environmental certification. I designed and developed the entire application, fully optimized it for SEO with advanced configuration for Google to maximize visibility, and integrated smooth animations to ensure a modern, fast, and fluid user experience.

Next.js
TypeScript
i18n
React
TailwindCSS
Framer Motion
Elsan

Elsan

I collaborated with the Void team to develop the Elsan web app. Elsan is one of the leading private healthcare providers in France. The project aimed to completely redesign their corporate website to meet new requirements for performance, modernity, and web visibility. I was responsible for conception to implementation, including requirements analysis, architecture design, front-end development with Next.js, and back-end development with Drupal 10. This large-scale project handles approximately 8 million views per month.

JavaScript
PHP
Drupal
Next.js
MySQL
TailwindCSS
Docker
Git
Drush
Monorepo
Greencore

Greencore

I am the Lead Tech Engineer behind Greencore, where I handle the architecture, conception, and coding. Greencore is a startup accelerating the energy transition through smarter projects and financing. We help projects in the energy sector secure financing while enabling companies to monitor, optimize, and reduce their energy consumption with AI-powered insights. Join us in driving the sustainable revolution.

PHP
TypeScript
Laravel
React
Inertia
TailwindCSS
PostgreSQL
Shadcn UI
Portfolio for an Artist

Portfolio for an Artist

I developed a professional portfolio website for Mourad Boulhana, a renowned Moroccan artist, director, and cinematographer. Known for his expertise in commercials and documentaries, the portfolio was designed to showcase his diverse projects from documentary films to experimental works highlighting his unique ability to capture fleeting moments of life and human interaction. This project aimed to create a modern and engaging digital presence that reflects his artistic identity and vision.

Next.js
TypeScript
React
TailwindCSS
Framer Motion
Vercel
Poultry Farm Management System

Poultry Farm Management System

Developed a comprehensive poultry farm management system enabling full oversight of farm operations. The application organizes data hierarchically from farms to batches, buildings, weeks, and daily records, allowing detailed tracking of mortality, feeding, treatments, weights, and analyses. It features automated calculations, cumulative totals, sequential validations, and weekly progression controls to ensure data accuracy and operational efficiency, supporting informed decision-making for farm management.

Rust
Tauri
TypeScript
React
TailwindCSS
Shadcn UI
Vite.js
SQLite
Geema

Geema

Freelance project for the development of a CMMS (Computerized Maintenance Management System) in collaboration with a branch of Cossona Enterprise. The project aimed to reduce machine downtime and optimize maintenance processes. This desktop application provides comprehensive maintenance scheduling, equipment tracking, and performance analytics to streamline industrial operations and improve efficiency.

TypeScript
Electron.js
React
Shadcn UI
TailwindCSS
MySQL
Association Management App

Association Management App

The Association Management App is a powerful web application designed to streamline the management of association members and provide a comprehensive set of features for efficient administration. This full-stack solution enables organizations to manage member databases, track activities, handle communications, and automate administrative processes with an intuitive user interface.

JavaScript
PHP
React
Laravel
TailwindCSS
Shadcn UI
ViteJS
MySQL
Licences et certifications

Professional Certifications

I've completed specialized courses and certifications to enhance my skills in artificial intelligence and machine learning, staying current with the latest technologies and best practices.

  • A

    Advanced Learning Algorithms

    Stanford Online

    Completed advanced coursework in machine learning algorithms, focusing on neural networks, tree ensembles, and optimization techniques. This certification demonstrates proficiency in implementing and understanding complex AI algorithms.
    TensorFlow
    Artificial Neural Network
    Tree Ensembles
  • S

    Supervised Machine Learning: Regression and Classification

    Stanford Online

    Comprehensive certification covering supervised learning fundamentals including linear and logistic regression, gradient descent optimization, and regularization techniques to prevent overfitting.
    Supervised Learning
    Logistic Regression for Classification
    Linear Regression
    Gradient Descent
    Regularization to Avoid Overfitting
    Python
Contact

Get in Touch

Want to chat? Just shoot me a dm with a direct question on twitter and I'll respond whenever I can.