AI/ML, Deep Learning Tech Talks

Important Note: Register on AICamp website is REQUIRED for admission.

Description:
Join the Capgemini, Qualcomm, Arduino, and Moorcheh for a discussion on the future of Agentic AI beyond the datacenter. We'll explore how advances in edge computing, local inference, vector search, and persistent memory are enabling intelligent systems to operate directly on devices while maintaining context, autonomy, and responsiveness.

The evening will combine enterprise perspectives, technical deep dives, and live demonstrations showcasing what becomes possible when AI systems can remember, reason, and act without constant reliance on cloud infrastructure.

Agenda:
* ​5:30 PM – 6:00 PM: Check-in, food, and networking
* 6:00 PM – 6:15 PM: Welcome & community update
* ​6:15 PM – 8:00 PM: Technical Talks & Q&A
​* 8:00 PM – 8:30 PM: Open discussion, networking, and closing

Opening Keynote: The Future of Compute at the Edge
Speaker: Evgeni Gousev, Vice President, Tech & Product Exec and Board Member, Qualcomm.
Abstract: A visionary opening on the evolution of edge computing. Evgeni explores how Qualcomm’s architecture empowers devices like the Arduino UNO Q to process complex, localized AI workloads without relying on massive datacenter infrastructure.

Tech Talk: ​A Technical Walkthrough of the Arduino UNO Q Platform
Speaker: TBD, Senior developer, Arduino (a Qualcomm company)
Abstract: ​This session covers the exact hardware constraints, configuring the Linux environment, and the mechanics of optimizing the board to support heavy local deployments like semantic search and open-source models.

Tech Talk: ​​The Moor Edge: Moorcheh ITS Vector Search Engine on Arduino Uno Q
Speaker: ​Dr. Majid Fekri, CTO & Co-Founder, Moorcheh.ai
Abstract: ​​A live code review and hardware demonstration showcasing the Moorcheh Information-Theoretic Search (ITS) search engine running entirely locally. Majid will walk through the privacy-preserving retail kiosk pipeline, demonstrating how the hardware searches its persistent memory and generates real-time answers—entirely offline. By running a full vector search and quantized local LLM natively on the Arduino UNO Q (3.6GB), we are proving that enterprise-grade agentic memory doesn't require massive compute.

Tech Talk: ​​Gen Edge AI: Reasoning, Agentic Memory, and Physical Intelligence
Speaker: Tara Khani, CEO & Co-Founder, Moorcheh.ai
Abstract: ​​Tara will detail how resource-efficient agentic memory structures serve as the missing layer for physical AI. As AI workflows transition from static RAG to autonomous coding agents (like Cursor and Claude Code), traditional vector databases introduce a catastrophic "RAM Tax" and passive retrieval bottlenecks. This session covers the architecture of Memanto by Edge AI Innovations, an open-source active memory agent built on an Information-Theoretic semantic engine. We will break down how to replace passive data ingestion with a three-primitive memory loop (remember, recall, answer) that gives agents persistent, cross-session context.

Speakers/Topics:
If you have a keen interest in speaking to our community, we invite you to submit topics for consideration: Submit Topics

Sponsors:
We are actively seeking sponsors to support our community. Whether it is by offering venue spaces, providing food/drink, or cash sponsorship. Sponsors will not only speak at the meetups, receive prominent recognition, but also gain exposure to our extensive membership base of 50,000+ AI developers in San Francisco and 500K+ in global.