AI

Description “Bro VS SiS Game” is an action-packed competitive game developed using Godot Engine. This project represents a complete remake of the original concept, featuring enhanced gameplay mechanics and improved graphics. The game focuses on sibling rivalry through fun and engaging competitive gameplay elements. Designed for Reinforcement Learning: The game is specifically designed to serve as a training environment for reinforcement learning models, allowing AI agents to learn how to play through interaction with the game environment. ...

Description Intelligent Agentic chatbot developed with LangChain for the “El Viejo Arancel” web platform. The chatbot was implemented using Gradio to create an interactive user interface and exposed to the web through FastAPI. The bot is deployed on Hugging Face and is currently part of the website’s paid section, providing an innovative way to assess product tariffs for users through an API. Technologies Used LangChain Python Gradio (User Interface) FastAPI (Web API) Hugging Face (Deployment) AI/Machine Learning Chatbot Development Natural Language Processing (NLP) Key Features Advanced natural language processing Intelligent contextual responses Payment system integration Cloud deployment with Hugging Face Agentic architecture for autonomous decisions Links Deployed on: Hugging Face (El Viejo Arancel payment system)

Description GroceryTracker is an innovative web application that demonstrates the power of multimodal AI models for practical everyday use. This project leverages the cutting-edge Mixtral Pixtral 12B 2409 model to automatically extract and organize grocery receipt information, transforming a simple photo into structured, actionable data. Built as a React application and deployed on Hugging Face Spaces, GroceryTracker represents an exploration into the capabilities of modern multimodal AI systems for real-world document processing and data extraction tasks. ...

Description Comprehensive automation project that leverages Large Language Models (LLMs) to evaluate SAS certification exams. This initiative resulted in the creation of a competitive leaderboard and the fine-tuning of a Llama 3.3 70B model using DeepResearch technology from Claude, covering all topics included in the SAS certification exam. Technologies Used Large Language Models (LLMs) Llama 3.3 70B Claude DeepResearch Hugging Face SAS (Statistical Analysis System) Machine Learning Model Fine-tuning Automated Evaluation Systems Key Features Automated Exam Evaluation: LLM-powered system to assess SAS certification knowledge Competitive Leaderboard: Real-time ranking system for exam performance Fine-tuned Model: Custom Llama 3.3 70B model trained on comprehensive SAS topics DeepResearch Integration: Utilized Claude’s DeepResearch for thorough topic coverage Comprehensive Coverage: Includes all SAS certification exam topics Performance Metrics: Detailed analytics and scoring mechanisms Project Outcomes Successfully automated the SAS exam evaluation process Created a transparent leaderboard system for competitive assessment Developed a specialized fine-tuned model for SAS-related queries Comprehensive knowledge base covering all SAS certification topics Links Hugging Face Organization