CTO interview: Roey Mechrez, changing the world of Artificial Intelligence
Roey has an impressive background and his credentials in the Artificial Intelligence world are proven. In 2018, he co-founded BeyondMinds, delivering hyper-customized, production-ready AI systems that enable sophisticated companies to overcome the massive failure rate in AI adoption.
You have a very impressive background, coming out of the Technion, the Israeli equivalent of MIT. It’s rare I interview a CTO who did a PhD. How influential was that in setting up BeyondMinds?
I see beyond BeyondMinds as a deep tech company. Our product and development are very much research oriented, and we’re out to solve one of the most complex problems with a huge market potential: Enabling easy production of AI solutions.
As a PhD in AI, I have a deep understanding of the problem, and the research required to solve it. Because we are a deep tech algorithmic AI driven company, that’s a critical factor.
On the other hand, I lack hands-on engineering experience, and to solve that we’ve built a great team working together around me. Without them, we would never be able to build a great product.
Non-techies might struggle to understand what BeyondMinds is doing, can you please elaborate in simple terms?
In simple words, bringing AI solutions to production, creating real value for enterprises, is very hard. It takes years and it’s highly expensive. Even once the solution is deployed in production, you need to invest a lot of effort to maintain and support it in the face of dynamic data and changing market conditions.
I heard of companies employing one Machine Learning engineer for each model in production. And that’s just for maintaining a solution in production!
The platform and product we build enable the productisation of AI solutions and maintenance them in production. The concept behind our product is a paradigm shift from model-centric development to a production-centric approach.
In model-centric development, you build one specific AI model and put a lot of effort to bring it to production. A production-centric approach, on the other hand, focuses on building all the components you need to productionize AI and prepare it to face real-world data environments of actual enterprises. This includes factors such as model robustness, sustainability, speed and ability to scale-up. With that framework in hand, you can onboard any desired AI models simply and quickly, enabling enterprises to embark on an AI transformation journey.
How does it compare to tools offered by AWS and GCP?
To answer that, we’d need to dive deep into the AI landscape; I’ll try to keep it simple and at a high level. We’re not building a tool for developers. Rather, we cater to line-of-business owners, who are not necessarily data scientists or AI experts. In fact, they don’t care too much about your technology: what matters to them is business impact and results that AI can help them achieve. Take defect detection in manufacturing plants, for instance. That’s actually one of the solutions that we build for manufacturing enterprises on top of our platform. The person using our system is an expert in the assembly line they manage, not in AI.
One of the most important components in our platform is the ability to collect feedback from users and retrain AI models once in production. That requires monitoring, retraining, dealing with data and many other elements that are part of the platform. We bring this solution to production and maintain them in production over time.
Why do you think companies want to adopt AI?
I think that AI is an amazing technology that enables a lot of things. In coming years, we will see more and more drastic changes and new value in terms of AI. If you look at Fortune 1000 companies, they invest, on average, $50 million a year in AI development. That’s because the potential value is enormous. Many of them realise that reaching profitability and long-lasting value from AI is extremely challenging.
At BeyondMinds, we partner with enterprises from a variety of industries: banking, insurance, and manufacturers, helping them obtain value from AI without being a Big Tech or having an engineering mentality.
How do you make sure you document your processes for growing and changing teams?
We are almost 70% tech people in the company, including research, engineering, and product. Knowledge sharing is done through code versioning and documentation. That enables us to maintain unified frameworks for documentation for developers. Personally, I’m a big fan of Notion, I think it’s an amazing platform.
What are the main challenges you find recruiting talent in the current employment market?
This is a challenge that everyone in today’s job market is facing, especially in the start-up community. We’re doing our best to maintain our great team, investing in our people and helping them develop professionally, as our company grows.
I deeply believe that companies hire talent so that these people will tell companies what to do, not the other way around.
If you want to connect with Roey, click here.
To learn more about BeyondMinds, visit their website: beyondminds.ai
If you’re a techie working on something exciting or you simply want to have a chat, get in touch with me. I’m currently CTO at Kolleno.com, working to modernise credit control teams.
