Projects focused on LLM-assisted reasoning, system control, and end-to-end intelligent workflows.

๐Ÿง  Alinea: Multi-Agent Group Intelligence

๐Ÿ‘‰ View Project Page

Alinea multi-agent demo screenshot

Description:
A multi-agent collaboration system built inside secure XMTP group chats. Instead of producing a single LLM response, Alinea orchestrates multiple AI personas โ€” Skeptic, Creative, Pragmatist, and Coordinator โ€” to deliberate in structured breakout threads before synthesizing a final recommendation.

The system integrates XMTP Agent SDK, Convos, and Kilo Gateway to explore humanโ€“AI collaboration in encrypted, multi-user environments. Designed with deterministic orchestration logic and fixed-round convergence for demo stability.


๐Ÿค– AI-Chatbot (LLM-Assisted Conversational System)

๐Ÿ‘‰ View Project Page

PartSelect Parts Assistant screenshot

Description:
An AI-powered conversational appliance repair assistant focused on dishwashers and refrigerators.
The system uses deterministic dialog routing with explicit state machines, and selectively applies an LLM only as a fallback parser for ambiguous natural language โ€” preventing chatbot drift and ensuring reliable troubleshooting.

Supports guided diagnostics, part compatibility confirmation, installation guidance, and demo order / human handoff flows.


๐ŸŽฎ Nexus: Full-Stack Esports Analytics Platform

๐Ÿ‘‰ View Project Page

Nexus Esports Analytics Platform screenshot

Description:
An end-to-end full-stack analytics platform for professional esports, enabling interactive exploration of Valorant and League of Legends data across players, teams, maps, agents/champions, and match performance.

Built with a React frontend and Node.js + Express backend, backed by SQL-based analytics queries over structured esports datasets.



Table of contents