A fully agentified extension of the Design2Code workflow with multi-agent LLM orchestration. Coordinates White and Green agents for HTML/CSS generation and evaluation using Playwright-based rendering, IoU geometric similarity, and multi-device screenshot comparison.
A multimodal pipeline that transforms mobile, tablet, and desktop wireframe sketches into clean, responsive HTML/CSS. Features vision-enabled LLM interpretation, multi-breakpoint reasoning, and comprehensive evaluation suite.
A domain-specific recommendation engine that matches PhD proposals to the most relevant faculty advisors. Built with LlamaIndex, LangChain, LangGraph, Qwen models, and ChromaDB for expertise extraction and relevance scoring.
A scalable platform for wearable-device research, providing participant management, real-time data ingestion, interactive dashboards, and AI-ready time-series pipelines. Used in behavioral science and health research.
A robust, researcher-friendly platform for behavioral studies. Supports a large library of experiments including Ultimatum Game, Dictator Game, Memory tests, Task Switching, Go/NoGo, N-Back, BART, Stroop, and many others. Used in peer-reviewed psychology research.