Omar AbedelKader

PharoGPT - 2026

Pharo-Copilot is an AI-powered code-completion assistant integrated directly inside the Pharo IDE, designed to help developers write Pharo code more efficiently.
It is built on local lightweight language models, trained specifically on Pharo repositories to support a low-resource programming language with no existing benchmarks or models.
The project provides a full toolchain: dataset mining tools, a tree-sitter Pharo parser, custom benchmarks, and an IDE plugin that embeds the completion models.

An AI-powered coding assistant integrated directly into the Pharo environment.
Lets developers interact with LLMs to explore, inspect, and query the live image.
Supports function calling so models can safely access packages, classes, and methods.

INERIS-IA is a tool designed to help CIVS engineers manage and analyze textual documents by integrating several NLP and text-mining functions.
It embeds pretrained ML models to automatically classify documents according to INERIS strategic goals (COP).
It provides a simple graphical interface built with Flask + MVC, allowing non-technical users to load models, run analyses, and explore text-processing features.