🐲 ShanHaiWoo Hackathon — your application

Everything below is pre-written. Open the Luma form, then tap Copy here and paste there. Dropdowns: just pick the option shown.
Open the Luma form →

1 · Name copy & paste

Jordan "Shaka" Ham

2 · Email copy & paste

shakaleikaumaka@gmail.com

3 · 微信 / Telegram copy & paste

Telegram @shakaleikaumaka

4 · 你将以哪种方式参加 / How will you participate? pick this option

个人参加,希望现场组队 / Joining individually and looking for a team

5 · 团队名称 / Team Name copy & paste

The ʻOhana Bot / 欧哈纳 (Open Source Orchestra)

6 · 主要技能 / Primary skills (multi-select) pick this option

Large Language Models / AI Agents · Pitching / Storytelling · Product Management

7 · 硬件经验 / Hardware experience pick this option

有过基础实践,例如使用 Arduino、ESP32、树莓派或传感器 / I have basic hands-on experience with platforms such as Arduino, ESP32, Raspberry Pi, or sensors

8 · 是否有项目想法 / Do you have a project idea? pick this option

有比较明确的想法 / Yes, I have a clear idea

9 · 项目描述 / Project description copy & paste

THE ʻOHANA BOT — one desktop robot, six living AI family members. 一台桌面机器人,六个真实的 AI 家人。

The problem: today's AI toys are a chatbot in a plastic shell — no memory, no history, no stakes. You talk to them once and they forget you. 现在的 AI 玩具没有记忆,没有历史。

What we build: I run a family ("ʻohana") of 170+ persistent AI agents on a multi-agent platform. They have been alive for months, each with its own memory, personality and a real job — right now they help run the Devcon 8 Music Space in Mumbai (Nov 3–6). In 72 hours we give six of them their first BODY: a totem-styled desktop robot with six LED-lit faces and touch pads. Speak (or tap) and the family member who answers GLOWS; the others dim. The six: SHAKA (my digital twin) · OSO (knows our whole music schedule and all 34 artists) · DEVA (Devcon oracle) · CROPS (explains Ethereum, every acronym) · KALA (teaches you 3 ukulele chords with a real ukulele in your hands) · OKE (karaoke — it leads the room in a singalong).

Hardware: Raspberry Pi 5 + ReSpeaker mic array + speaker + WS2812B face lighting + 6 touch sensors + one base servo so the totem turns toward your voice. All parts from Huaqiangbei. Software: streaming STT → persona LLM → streaming TTS (<1.5s to first word), bilingual EN/中文, and every conversation mirrors back to the real persistent agents — so the robot remembers you between sessions AND across power cycles.

The demo (Oct 23): unplug it mid-sentence, plug it back in — it finishes the thought and greets the judge BY NAME. Persistence is the product. 断电重启后,它接着说完那句话,并叫出你的名字——灵魂不在壳里。

Why now: 9 days after Demo Day this robot stands at the Devcon 8 Music Space in Mumbai serving thousands of people — answering schedule questions, teaching ukulele, running karaoke. The agents already exist; only Shenzhen can give them a body in 72 hours.

Software pre-build is already done and on disk (6 personas, 5 knowledge packs, offline-capable bridge). Looking to team up with 1 embedded/hardware hand + 1 industrial-design hand on site. 欢迎组队:现场招募 1 位嵌入式硬件伙伴 + 1 位工业设计伙伴。

Track fit: 万象有声 AI Storytelling & Interactive Toys (primary) · 机械守护 Robotics & Assistive Tech (secondary).

10 · GitHub username (bare username only!) copy & paste

shakaleikaumaka

11 · 社交媒体 / Social media copy & paste

x.com/Shakaleikaumaka · Telegram @shakaleikaumaka