Junxiao Guo

Projects

MonoLLM

MonoLLM

A powerful framework that provides a unified interface for multiple LLM providers, allowing developers to seamlessly switch between different AI models while maintaining consistent API interactions.

Go to project
Omnichat - All-in-One LLM Chat Application

Omnichat - All-in-One LLM Chat Application

A modern chat application for interacting with multiple AI models from different providers. Available as both a web app and desktop application.

Go to project
LLM Agent Workflow Generator

LLM Agent Workflow Generator

A minimal agentic LLM web app built with Next.js 15 (App Router), featuring a chat interface and a visual workflow editor for AI agent flows.

Go to project
Company Website Template - Next.js

Company Website Template - Next.js

This is a template for building a modern company website using Next.js (App Router), TypeScript, Tailwind CSS, and shadcn/ui. It includes features for managing content like blogs and news articles via Markdown files.

Go to project
SimpleLM: Lightweight Transformer Language Model with RLHF (PPO)

SimpleLM: Lightweight Transformer Language Model with RLHF (PPO)

SimpleLM is a minimal, CPU-friendly language model based on a Transformer decoder architecture. It supports both supervised fine-tuning (SFT) with prompt/response pairs and PPO-based RL training with a neural reward model (RLHF-style) for human/AI feedback (RLHF) using PPO. The project is designed for educational purposes and easy experimentation on small hardware.

Go to project
DayRider

DayRider

Simplify your weekly work with a minimal todo task management app.

Go to project
An End-to-end Image Dehazing System with Light-weighted CNN And Dark Channel Prior

An End-to-end Image Dehazing System with Light-weighted CNN And Dark Channel Prior

Instructor: Joseph Konan Team: Jingsi Chen,Junxiao Guo, Shuhao Ren ,Luo Yu (ordered by last name initials)

Go to project
Swarm

Swarm

A comprehensive CLI-based agent for web browsing, automation, and deep research with LLM integration.

Go to project