Cupertino AI Meetups

May 27, 2026
9:00 AM - 11:00 AM PST
Online. Register for Zoom!

Video-Native Multimodal Models for Video and Image Understanding

In this 20-minute talk, Akshat will introduce Perceptron’s latest release, a video-native multimodal model that matches or exceeds frontier models from Google and Alibaba on video and image understanding at a fraction of their inference cost. He’ll walk through the capabilities that move the needle for real video workloads: temporal grounding to clip precise events from long streams, egocentric reasoning for first-person and wearable contexts, and structured “thinking traces” that reason over motion and physical space. He’ll also cover the image-side advances production perception teams care about: reliable pointing, point-by-example one-shot visual search, dense counting, dial/gauge/clock reading, and structured document extraction.

About the Speaker

Akshat Shrivastava is the CTO and co-founder of Perceptron, previously leading AR On-Device at Meta and conducting research at UW.

Getting Started with Perceptron AI in FiftyOne

In the second half of the session, Harpreet Sahota will walk through how to get started using Perceptron’s video-native multimodal model within FiftyOne for real-world video understanding workflows. He’ll demonstrate how to connect to the API, explore multimodal outputs inside FiftyOne, and build practical workflows for tasks like temporal event analysis, visual search, and video dataset inspection. Attendees will leave with a hands-on understanding of how to integrate state-of-the-art video perception models into their existing computer vision pipelines.

About the Speaker

Harpreet Sahota is a hacker-in-residence and machine learning engineer with a passion for deep learning and generative AI. He’s got a deep interest in VLMs, Visual Agents, Document AI, and Physical AI.

Jun 09, 2026
9:00 AM – 10:30 AM PST
Online. Register for the Zoom!

Medical imaging teams are increasingly fine-tuning foundation models like UNI, MedSAM2, and BiomedCLIP on small in-house datasets. At that scale, label disagreement is a dominant cause of model failures, and the disputed ground truth is what regulators will ask you to defend. We'll build a medical imaging dataset in FiftyOne, surfacing and analyzing the cases where reviewers disagree. From there, we'll fine-tune a foundation model on cleaned data and use FiftyOne to evaluate where our model succeeds and fails, and which data is needed to move the model’s performance forward.

You’ll learn how to:

• Build a medical imaging dataset that preserves multiple expert annotations as first-class fields
• Use FiftyOne views, embedding similarity, and confidence-disagreement signals to find the samples where reviewers split.
• Run label-quality screens, near-duplicate detection, and active-learning sample selection using foundation model embeddings
• Fine-tune a medical foundation model on a defensible dataset, with auditable and versioned experiment tracking
• Filter and slice evaluation for regulatory and clinical readiness
• Drive the pipeline with natural-language agents using the FiftyOne MCP Server and Skills to run the same curation, evaluation, and review workflows from your favorite AI tool

Who This Is For

• ML and computer-vision engineers in the medical imaging space
• Data and annotation operations teams
• Clinical AI and digital pathology leads
• Regulatory and quality leads

Jun 17, 2026
9 AM Pacific
Online. Register for the Zoom!

AI agents are rapidly changing how teams build and scale machine learning workflows—but most implementations still rely on fragmented tools, manual processes, and brittle integrations.

We’ll walk through how modern agentic systems move beyond simple prompts—leveraging structured tools like dataset operations, embeddings, evaluation pipelines, and model execution to take real action. You’ll see how these agents can tag data, run inference, evaluate performance, and surface insights automatically, all within a unified workflow.

By combining natural language interfaces with programmable building blocks, teams can dramatically reduce manual effort, accelerate experimentation, and unlock faster decision-making across the ML lifecycle.

Whether you're building data-centric AI systems, managing large-scale vision datasets, or exploring agentic workflows for the first time, this session will give you a practical blueprint for getting started.

What you’ll learn:

• How AI agents actually work in production: Move beyond hype—understand how agents use tools, memory, and structured workflows to execute real tasks
• Using tools to take action on your data: Program agents to run operations like filtering labels, computing embeddings, evaluating detections, and more
• What “skills” are and how they enable multi-step workflows: Learn how to package complex logic into reusable capabilities your agents can call on demand
• How MCP connects your models, tools, and agents: See how MCP standardizes communication between LLMs and external systems for scalable, flexible architectures
• Automating the ML lifecycle with agents: From data curation to model evaluation, discover how to eliminate repetitive workflows and accelerate iteration
• Best practices for building reliable agentic systems: Design patterns, guardrails, and practical tips for deploying agents in real-world environments

This workshop explores the FiftyOne plugin framework to build custom computer vision applications. You’ll learn to extend the FiftyOne App with Python based panels and server side operators, as well as integrate external tools for labeling, vector search, and model inference into your dataset views.

Time, Date and Location

Jun 24, 2026
9 AM - 10 AM PST
Online. Register for the Zoom!

What you'll learn:

• Build Python plugins. Define plugin manifests and directory structures to register custom functionality within the FiftyOne ecosystem.
• Develop server side operators. Write functions to execute model inference, data cleaning, or metadata updates from the App interface.
• Build interactive panels. Create custom UI dashboards using to visualize model metrics or specialized dataset distributions.
• Manage operator execution contexts. Pass data between the App front end and your backend to build dynamic user workflows.
• Implement delegated execution. Configure background workers to handle long running data processing tasks without blocking the user interface.
• Build labeling integrations. Streamline the flow of data between FiftyOne and annotation platforms through custom triggers and ingestion scripts.
• Extend vector database support. Program custom connectors for external vector stores to enable semantic search across large sample datasets.
• Package and share plugins. Distribute your extensions internally and externally