logo

Every dataset has a story waiting to unfold 


Resume Parser

An advanced resume parsing system leveraging LLMs to extract and analyze key information from resumes. The tech stack includes:
  •   FastAPI: For building a high-performance, easy-to-use API
  •   OpenAI API: To enhance text analysis and information extraction
  •   Next.js: For creating a responsive and dynamic front-end interface
  •   Redis: For caching and improving response times
  •   Docker: For containerization and easy deployment
The system employs OpenAI API to accurately extract information such as contact details, work experience, education, and skills from various resume formats.

Assistants App

A Retrieval-Augmented Generation (RAG) platform enabling users to create, customize, and interact with AI assistants. Key features include:
  •   Assistant Creation: Users can create multiple AI assistants tailored to specific tasks or domains
  •   File Integration: Ability to upload and associate files with each assistant, enhancing their knowledge base
  •   Interactive Chat: Engage in conversations with created assistants, leveraging their specialized knowledge
The tech stack comprises:
      LangChain, Milvus, Ollama, OpenAI API, FastAPI, Nextjs, PostgreSQL, ...

Invoice OCR

A robust system for extracting key information from invoices through a combination of Visually Rich Document Analysis, OCR and NLP. Invoices are often complex documents with varied layouts, fonts, and graphical elements, making automated data extraction challenging. For the api, I used Fastapi, Celery, Postgresql, SQLalchemy, RabbitMQ, Redis, ...
UFC Dataset

UFC Dataset

Explore the Ultimate Fighting Championship (UFC) with this comprehensive dataset. The dataset includes granular fight results, fighter profiles, and referee insights. Data is scrapped using Scrapy and constantly updated using Airflow and the Kaggle API, ensuring it remains current and accurate for in-depth analysis and research.

KYC Application

A KYC application designed to extract data from images of passports (any) and two types of Moroccan identity cards (old and new) using computer vision models and OCR technology.

Bookstore

A bookstore application built with PHP Symfony framework. It has a user authentication system, CRUD operations, a search feature, an admin dashboard ...