fullstack-ai

Advanced Digit Recognition System using a Hybrid Modern-CNN (ResNet + ConvNeXt) & Ensemble Inference Engine. Features a robust OpenCV preprocessing pipeline and FastAPI backend.

Institutional-grade Trading Terminal. Next.js 14, GPT-4 RAG Architecture, Microsoft Qlib ML Signals, & Real-time Market Telemetry. Full-stack AI Ecosystem.

Manage AI agents to add structure, memory, and accuracy to your AI-driven development projects across full stack applications.

Full-stack AI semantic search platform built with FastAPI, React, MongoDB, and SentenceTransformers.

An Emotional AI Companion built with FastAPI & Groq. Features real-time Personality Analysis (Roman Urdu + English), AI Image Generation, and Mood Analytics. Designed by an AI Enthusiast to explore advanced Human-AI interaction.

This is an AI-powered waste classification application that uses computer vision to detect and categorize different types of waste materials.

Lightweight Voice-enabled RAG assistant built with FastAPI, Groq LLM, and Google Generative AI. It performs document-based question answering with speech input/output support and is fully containerized using Docker for easy deployment.

Production-Grade AI Engineering Portfolio by Muhammad Ibrahim: Generative AI, GraphRAG, Multi-Agent Swarms, AI Safety Guardrails, MLOps, Computer Vision (YOLOv8/VLMs), Privacy-Preserving AI & Full-Stack AI Mobile/Web Apps.

Engineered a containerized backend service using Docker and FastAPI, focusing on resource efficiency and deployment portability.

This is an AI-powered waste classification application that uses computer vision to detect and categorize different types of waste materials.

Full-stack AI system for plant disease detection using YOLOv8, CNN, U-Net, and LLM (Cohere + RAG) with real-time analysis and dashboard.

Cross-platform AI agriculture application built with React Native and Flask. Features Scikit-Learn crop recommendation algorithms and a custom PyTorch ResNet9 computer vision model for real-time plant disease classification.

An asynchronous full-stack RAG application built with an Express.js API gateway, Redis task runner, and a Python LangChain engine to analyze, index, and chat with public GitHub repositories.

From vibe coding to full-stack AI development. Drop one folder into any project - structured planning, persistent memory, skills, and zero drift for any AI coding tool.

Nova AI — a full-stack AI chatbot built with React and Express featuring smart follow-ups and developer-friendly responses.

Production Full-Stack RAG Recommendation Engine | Spring Boot 3 + PostgreSQL pgvector (HNSW) + Gemini LLM + Streamlit Geospatial UI

Introducing Bimo 5 (Autonomous Multi-Modal AI Agent with real-time voice, vision processing, document analysis, image generation, and web search/fetch).

This project aims to provide a robust, end-to-end framework for detecting financial fraud by combining traditional pattern recognition with behavioral anomaly detection. By bridging the gap between a high-performance machine learning backend and a modern observability dashboard, it offers a holistic view of transaction safety in real-time.

An advanced, enterprise-grade AI security system combining real-time computer vision (YOLOv8) with a local Multimodal RAG architecture (FastAPI, ChromaDB, and Llama 3 via Ollama) to allow natural language querying over recorded surveillance logs.

Universal Multi-Agent Task Orchestration & AI Decision Ecosystem (Jev System 1 Routing, Gemini Reviewers, Blackboard Memory)