HackBuddy: Building an Open-Source AI Event Concierge After Receiving 500+ Hackathon Registrations

Описание и применение

HackBuddy – это открытый AI‑консьерж для мероприятий, который автоматизирует ответы на вопросы участников и предоставляет организаторам аналитические данные, используя локальные модели через Ollama и интеграцию с Arduino.

Задача

Community event organizers faced repetitive attendee questions and limited capacity, making it hard to manage over 500 registrations for Hacktoberfest Hack Day Mumbai.

Результат

HackBuddy streamlined attendee communication, reduced organizer support time, and enabled data‑driven insights for a large hackathon event.

Рабочий процесс

• 1. Collect event data as Markdown files (FAQ, schedule, venue, sponsors, challenges).
• 2. Ingest Markdown into a searchable knowledge store.
• 3. Deploy local LLMs (Gemma, Qwen, Llama) via Ollama for cost‑effective inference.
• 4. Use Microsoft Agent Framework to orchestrate interactions between knowledge store, recommendation logic, and sponsor discovery.
• 5. Integrate Arduino UNO Q to display status LEDs and enable future physical interactions.
• 6. Deploy the .NET/Blazor Server application locally for organizers.
• 7. Provide attendees with AI concierge for FAQs, project guidance, and sponsor discovery.

Доказательства

HackBuddy is an open-source AI event concierge built with .NET, Microsoft Agent Framework, Ollama, and Arduino.

Источник

As our event registrations crossed 500, it became obvious that many of these interactions could be streamlined through a dedicated event AI assistant.

Источник

HackBuddy acts as a digital event concierge that helps both attendees and organizers.

Источник

Event-specific information is stored as Markdown files.

Источник

Instead of relying on expensive cloud APIs, HackBuddy runs against local models through Ollama.

Источник