Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Secure AI Training on Private Data
Learn how to build a decentralized, TEE‑based, encrypted data access system with blockchain, enabling secure private data aggregation for AI training, with code examples.
This session will provide a technical deep dive into building secure data access infrastructure that enables AI systems to train on private data without compromising privacy or security. We’ll walk through the complete architecture of a decentralized data access system that uses Trusted Execution Environments (TEEs), encryption, and blockchain technology to create a trustless environment for data access.
You’ll leave with practical knowledge of the technical challenges in private data aggregation and working code examples you can adapt for your own projects.
Vana is an open network for user-owned data, powering an AI economy.
This service transforms encrypted data via Docker, uploads to IPFS, and registers on-chain.
Related projects
Lessons from Deploying a Lightweight AI Consumer App: Turning AI Storybooks into Physical Products
Waterloo
This talk covers challenges integrating generative AI storybooks with print-on-demand services to create physical products and explores future…
AI-powered linter
Waterloo
This talk covers an AI-powered linter that lets tech leads easily write rules to enforce code quality, preventing…
LLM Fingerprinting: Identifying AI Models by Their Responses
Toronto
This talk demonstrates a system to identify and classify large language models by analyzing their responses to benchmark…
Leveraging Private AI with Model Context Protocol
Seattle
Overview of Kamiwaza's private AI platform, demonstrating a private chat with a private model and Model Context Protocol…
Enriching Global Trade Data with Deep Research
Waterloo
Learn how to automate creation of detailed company profiles from Bills of Lading using web scraping and LLMs,…
All the Trainingz, No Codez
Toronto
This talk demonstrates how to train, finetune, and preference tune large language models on a home computer using…