# Seattle Spark Hack Winners Livestream Spotlight: LiveKit & Memo

## Executive summary

This summary covers the NVIDIA DGX Spark Hack winners spotlight, featuring two local AI applications: LifeKit, an offline survival companion, and Memo, a local-first wearable assistant. Both projects demonstrate the capability of running complex, multi-modal AI workflows—including vision, speech, and retrieval—on edge devices like the Acer Veriton GN100. The core technical focus is on achieving reliable, low-latency AI functionality without constant internet connectivity, utilizing advanced NVIDIA models like Neotron and Cosmos.

## Key takeaways

- Offline AI Capability (LifeKit): LifeKit addresses the critical need for stable information access when internet connectivity fails. It uses local AI and Retrieval Augmented Generation (RAG) against extensive, downloadable knowledge bases (e.g., Wikipedia) to provide guidance on survival topics (first aid, navigation) without relying on external servers.
- Selective Memory and Privacy (Memo): Memo is a wearable assistant designed for elderly care, focusing on privacy by only tracking objects explicitly chosen by the user (e.g., keys, wallet). It uses visual evidence and confirmation to save location data, ensuring that the system does not record or store general video feeds.
- Concurrent Model Orchestration: Both projects rely on running multiple NVIDIA models concurrently (e.g., Neotron, Cosmos, Parakeet, VSSs, NIM, VLM) on the GN100. The primary architectural challenge is orchestrating these disparate tools—from speech transcription to reasoning—into a single, reliable, and deterministic application.

## Technical details

- LifeKit Architecture: The system utilizes Retrieval Augmented Generation (RAG) to source information from a locally downloaded, extensive library. Neotron 3 Nano acts as the default chatbot, ensuring quoted passages are verbatim and easily traceable. The system is designed to cue for online retrieval when offline, providing a response and then waiting for connectivity to source better information.
- Memo System Design: Memo runs on smart glasses, streaming camera and microphone data to the GN100. It uses Neotron 3.5 Lightning as the agent brain to route requests (visual memory lookups, questions, support requests). The system uses Cosmos 3 Nano to reason over short sequences of frames, allowing it to detect and verify an object's location change (e.g., keys being placed down) before saving it to memory.
- Edge Computing and Hardware: Both applications are built to run locally on the Acer Veriton GN100, powered by the NVIDIA GB10 Grace Blackwell Superchip. This architecture enables complex, multi-modal AI processing (vision, speech, reasoning) to occur entirely on the device, minimizing latency and dependency on external networks.

## Practical implications

- Enables critical services (e.g., emergency navigation, elder care) in areas with unreliable or zero internet access.
- Reduces privacy risks associated with cloud-based AI by processing sensitive data (video, location) locally on the device.
- Pushes the boundaries of wearable technology by integrating complex, multi-model AI into consumer-facing smart glasses.

## Topics

Local AI, Edge Computing, Wearable Technology, Agentic Systems, Retrieval Augmented Generation (RAG), Computer Vision, NVIDIA GB10 Grace Blackwell Superchip, Acer Veriton GN100, LifeKit, Memo, Neotron 3 Nano / Neotron 3.5 Lightning, Cosmos

Source: https://www.youtube.com/watch?v=QdR42yFeoQY