Pose Estimation, Milvus, Raspberry Pi, AI Kit, Edge AI, Edge Vectors, Images, Slack, minio, Python, Hailo 8L
Check in: We will open the door at 5:30PM for checkin, food, and networking. Please note that ALL guests must pre-register, as there will be no walk-ins or plus-ones allowed. Entry will be granted on a first-come, first-served basis, and once we reach capacity, the venue will not permit additional guests.
Agenda: The tech talks will start at 6pm. Check the event website for detail agenda.
The venue: Microsoft NYC — Times Square, 11 Times Square, New York, NY 10036. Room Name: Central Park West 6501
A big thanks to AI Camp and all their amazing organizers, they are an amazing team and cool people. Also thanks to all the speakers, great talks.
The first talk was by Rob Cheung from Substrate and was really interesting deep dive into how to process modern multi-model workloads. Mixture of Agents (MoA) was mentioned and is super interesting! Substrate is definitely something to check out.
Tech Talk: Agentic and Compound AI Done Right
Speaker: Rob Cheung (Substrate)
Abstract: This talk will cover the ins and outs of running multi-model workloads at scale. Learn about the latest techniques and frameworks to ensure your AI programs run fast and efficiently. Participate in a hands-on workshop that showcases practical applications of the concepts discussed in the talk. Gain insights into optimizing your AI workflows and see real-world examples in action.
Tech Talk: Human Pose Estimation in Real-Time Utilizing Edge AI Accelerated Hardware
Speaker: Tim Spann (Zilliz)
Abstract: Utilizing a Hailo AI Acceleration Module with a Raspberry Pi 5 device we will process real-time video streams from an edge camera and store real-time results in a vector database and send messages to Slack channels. We will show you how to build Edge AI applications that can stream unstructured data to the cloud or store it locally in a local Vector database. We will run a live demo of a neural network inference accelerator capable of 13 tera-operations per second (TOPS).
Pose Estimation and Milvus with Raspberry Pi 5 + Hailo 8L RPI AI Kit
Milvus, Slack, Python, Hailo, YOLO 8, Object Detection, Raspberry Pi AI Kit, Raspberry Pi 5 with 8GB RAM, logi webcam, Hailo8L M.2 module, M.2 M-Key Hat, Heat Sink.
So we take the example Pose Estimation code provided by Hailo team and add Milvus insert (cloud, docker or lite), Slack, Minio and REST image upload. Pretty easy.
AI Camp New York
Raspberry Pi 5 — AI Kit
Edge Vector Olympics
- Gold — NVIDIA Jetson AGX Orin — 275 TOPS, 2048-core, 64 GB RAM
- Silver — NVIDIA Jetson Xavier NX, 21 TOPS, 384-core, 8 GB RAM
- Bronze — Raspberry Pi 5, 13 TOPS, 4-core, 8 GB RAM, 12-core VideoCore VII GPU
COMMUNITY
Great reports from the community
As always the attendees to AI Camp and all the ML/Data/AI/Unstructured Data/Vector Database meetups are awesome and in numbers I haven’t seen since the start of the Apache Spark revolution. There is a great interest and people are interested in learning. It is a very excited time and I can’t wait to have more speakers, topics and fun meetups and events to bring to New York, Princeton, BellWorks and beyond. I try to get as much online for the remote people.
Get Milvused!
Read my Newsletter every week!
For more cool Unstructured Data, AI and Vector Database videos check out the Milvus vector database videos here:
https://www.linkedin.com/company/zilliz/
https://www.linkedin.com/in/timothyspann/
