python

A deep learning project for detecting and classifying teeth as normal or cavity using YOLOv12m. The model is integrated into a web application for real-time dental screening. Designed to enhance early diagnosis accessibility, with future improvements planned for dataset expansion, annotation quality, and robustness to real-world conditions.

Detecting fake news articles by analyzing patterns in writing.

Local-first AI assistant — 9 specialized agents (code, web, debug, security…), 10M-token vector memory, multimodal (vision + local image generation), two-way voice (dictation + read-aloud), slash commands & codebase RAG, agentic VS Code extension, mobile relay via secure tunnel. 100% local with Ollama. Your data, your control.

Intelligent Image Gallery with Uploads, Deduplication, and Text-Based Search Using Vector DB Qdrant

AI-native NoSQL + vector database with built-in semantic search for RAG, AI agents & LLM memory — the async Python client. Self-hosted Pinecone alternative.

A personal AI knowledge assistant powered by RAG, semantic search, and document understanding for context-aware answers.

Build a full‑stack helpdesk system with autonomous multi‑agent AI for ticket resolution and escalation.

A-MEM Agentic Memory System - MCP Server for IDE Integration (Cursor, VSCode) | Dual-Storage: ChromaDB + NetworkX DiGraph with explicit typed edges | Based on Zettelkasten

⚡ Vectro: Lightning-fast embedding quantization. Hit 12M+ vec/s throughput (4.85× faster than FAISS C++) while drastically cutting memory footprint for vector databases and local AI research.

An AI-powered quote generator bot that posts beautiful quotes to various social media platforms!

OpenViking knowledge base integration for nanobot — RAG augmentation, semantic search, vector embeddings, document indexing. Free models on SiliconFlow.

Local AI server with persistent memory, RAG, and multi-backend inference (MLX / llama.cpp / Ollama). Runs entirely on your machine — zero data sent to external services.

Experimental web console for real-time multimodal AI interactions (voice/video) using Gemini 2.0 Live API and NijiVoice. Next.js (TypeScript, Tailwind) frontend and FastAPI backend communicate over WebSocket for low-latency streaming.

EduMIND — Multimodal Bilingual Lecture Assistant with Anti-Forget RAG, Whisper ASR, VietMix NMT (CMI-aware), and Human-in-the-Loop Active Learning via Label Studio. Built for code-mixed Vietnamese–English academic environments.

面向 A 股投研，针对分析时效差、逻辑不全及数据难量化痛点，自主构建以 Qwen2.5-72B-AWQ 为核心的全链路自动化投研系统。利用 LangGraph 编排多智能体协作流，融合 LoRA 垂直微调、MCP 工具链与 CoT+ReAct 架构，实现从行情舆情实时采集、产业链知识图谱多维推演到 Reflexion 反思审核的投研全闭环，每日 9:00 定时产出机构级专业研报推送飞书，显著提升投研决策自动化水平。

Automated comic cataloging tool that identifies issues directly from cover images using a vision-language model, then cross-references results with the Grand Comics Database and the ComicVine API to generate structured, high-confidence collection data with minimal manual entry.

Deep learning models(CNN, LSTM, BERT) for image and text classification task with Tensorflow and Keras