Quantuity Analytics Inc. reposted this
In today's competitive market, engineers face increasing challenges in developing embedded products with ML chips that have power efficiency and performance for useful edge AI applications. In this episode of tinyML TALKS with LIVE Q&A, we're going to talk to Andrew Wright from eFabless about how they are enabling the creation of custom ML products through custom chips optimized specifically for extreme edge ML deployments! We'll walk you through an end-to-end design workflow—from ML model development to silicon—using accessible tools. You'll learn how to overcome common barriers in edge AI chip design and bring innovative, power-efficient products to market faster. What You Will Learn: - Tackling Edge AI Challenges: Discover how to design custom MVL products and chips optimized for power efficiency and performance in edge AI applications. - Seamless Design Workflow: Explore a simplified development process using Efabless’ SoC templates and automated tools to accelerate your design cycle while focussing on the value-add of your ML feature premise. - Power-Saving Innovations: Learn to design chips that meet the low-power needs of battery-operated devices without sacrificing performance.
tinyML TALKS - Solving Edge ML Challenges with Custom Chips with Efabless!
www.linkedin.com
Excellent presentation! Clear insights and valuable takeaways – really enjoyed how you broke down complex concepts into practical, actionable steps.
1- What are some unique challenges you encounter when designing chips for tinyML applications compared to larger-scale ML?
3-How does Efabless address the cost challenges in producing custom chips for tinyML, especially for startups or small companies?
2- How compatible are these custom chips with existing tinyML frameworks, and are there specific tools required for integration?
4- Do you plan any contests for universities?
In AI model specific Silicon, what are the important differentials from HW side?
Silicon Valley, ex-Microsoft, and Independent leader in Edge AI
1moWe’re going from CPUs to GPUs to NPUs to now customer silicon per model…tune in!