Projects

A curated selection of research and engineering projects spanning AI, full-stack development, and mobile applications.

AI & Research

ResearchRadar

An ArXiv intelligence tool that goes beyond listing papers. Automatically tracks multi-week citation trends, clusters papers into topic maps, and answers research questions via local RAG. Supports Ollama, DeepSeek, and OpenAI. All data stored locally.

Python TypeScript FastAPI RAG React

ReviewLLM

End-to-end pipeline for automatic academic peer review generation. Integrates retrieval-augmented prompts, reviewer persona conditioning, and rubric-aligned scoring to boost review helpfulness.

Python LLM NLP RAG

DaGAN Talking Face Reproduction

Reproduced the CVPR 2022 paper Depth-Aware GAN for Talking Head Video Generation. Trained on 4× RTX 4090 GPUs and evaluated with PSNR, SSIM, FID, LSE-C and LSE-D metrics. Fully Dockerized.

Python PyTorch GAN Docker

Full-Stack Applications

心迹 · Diary

A privacy-focused personal diary web app with rich-text editing, mood tracking, tag classification, calendar check-in, and writing statistics. Supports Docker deployment and data export.

React Node.js PostgreSQL Docker

Healthcare Platform (v2)

Upgraded full-stack health management system. Features appointment scheduling, health data tracking, medication reminders, community forum, and real-time WebSocket notifications. AI consultation via ERNIE Bot API.

Flutter Vue.js Spring Boot Redis

Healthcare Platform (v1)

First-generation health management platform built with Spring Boot and Vue.js. Provides health records, exercise tracking, diet management, drug lookup (OpenFDA), and food nutrition data (FatSecret API).

Java Vue.js Spring Boot MySQL

Mobile Applications

高考志愿填报建议系统

A Flutter mobile app that helps Chinese high school students select universities. Integrates college search, admission score history, AI-powered recommendation, and personalized score management.

Flutter Dart JWT REST API

What I am Looking For

I am eager to collaborate on projects that push LLM evaluation forward, especially around peer review and agent design. Feel free to reach out to explore ideas together.