This is the blog section. It has two categories: News and Releases.
Files in these directories will be listed in reverse chronological order.
News
International Roadmap for Devices and Systems
The semiconductor industry has a unique distinction among all the industries of having begun with a roadmap.
Many people are familiar with the idea of Moore’s Law and the pace of exponential growth in technology. Fewer understand that this is no accidental quirk of statistics, but instead due to significant pre-competitive global collaboration to deliver a rolling 15-year roadmap.
The IEEE is host to the International Roadmap for Devices and Systems and the 2023 update may be found at the link below:
Continuous Delivery Best Practices
The road to Continuous Delivery is not always a simple one, and has many a pitfall en route. There are resources that can help you navigate a smooth ride, however.
Our goal is not to build software, or models, but to build a product that leverages software or machine learning to solve the problem that the product addresses.
In many cases, this is a problem that has never been solved before, so there is no map to follow and no set of requirements to implement. Instead, we must learn the nature of the problem as we go, discovering many ways not to solve it before we find those that work.
To do so, we must work with customers and investors to find product-market fit, but can only do so within a limited budget of attention and resources. To succeed in this, we must optimise our processes towards iteratively discovering the product that meets our customer’s needs.
This is not the same as optimising our delivery for the convenience of the engineering team, which is a commonly held misconception.
The Continuous Delivery Foundation offers a helpful guide to best practices:
The Future of MLOps
Understanding the direction of MLOps capabilities
Typically, about 5% of the effort required to launch a product using AI is actual Machine Learning. Managing AI products in production involves navigating a complex morass of risks and challenges, many of which are non-obvious in advance of them becoming issues.
The MLOps Roadmap helps teams to understand the full scope of requirements in managing ML assets in production
