The integration of AI and big data opens new opportunities for the manufacturing sector. Discover in the interview how manufacturers ensure precise identification of every item or component using advanced analysis tools.
Can you share a bit about your background and journey that led you to become the CEO of Cybord?
Before joining Cybord, I was the VP of Supply Chain at Nvidia, and before that, VP of Operations at Mellanox Technologies. In these positions, I was responsible for the ownership of the entire supply chain process end-to-end, and consequently I witnessed a few quality-related issues associated with bad components that rapidly deteriorated into devastating recalls. After experiencing firsthand the importance of resolving poor quality components while the boards are still on the production line, I joined Cybord to help significantly reduce the risk exposure for Tier 1 OEMs like Nvidia and others.
What inspired you to start Cybord, and what is the mission and vision behind the company?
Cybord was founded by my colleague and CTO, Dr. Eyal Weiss, an industry expert who was previously the R&D department manager at a major defense solutions company. Following a crisis in a large-scale defense project due to faulty electronic components (the basic building blocks of an electronic circuit board), it took him and his team months to trace the problem down to a faulty capacitor worth three cents.
Following this incident and understanding the vast phenomenon of defective components, Cybord has undertaken a mission to remove these components from production lines. When I joined the team as CEO, I wanted to utilize my experience in supply chain and operations with large OEMs to help realize this goal and minimize manufacturing disruption by focusing on surgical traceability capabilities – the ability to pinpoint any problematic component and conduct rapid surgical recalls on an individual basis with minimal disruption to the manufacturing process – and electronic component health powered by AI and big data.
Cybord operates in the AI industry. How do you perceive the current AI landscape, and what challenges do you believe businesses face in this digital age?
The combination of AI and big data has introduced a new world of possibilities to the manufacturing industry. Traditionally, manufacturers have had to ensure that every individual item or component in the process was specifically identified by analysis tools. However, as the number of different items available on a typical production line is almost infinite, recognizing defects and traceability text content has been an impossible mission until now.
With the availability of robust AI and big data, the industry now has the opportunity to scale up. The digital age is generating huge amounts of data from the manufacturing process. The challenge is the ability to capture this data and sort it in a meaningful way so that production lines and OEMs can base their decisions on the data analysis results. AI is the perfect solution to harvest and analyze almost every possible piece of data from the production lines and generate value for the end customer.
The industry recognizes that whoever owns the data wins. Controlling the data enables manufacturers to gain a competitive advantage by making better decisions about how to allocate resources, optimize production lines, and identify defects early in the production process.
The result is a more efficient, productive, and customer-focused manufacturing industry. AI and big data are revolutionizing the manufacturing process, and the benefits are only going to expand in the years to come.
Could you provide an overview of your visual AI platform and how it ensures the reliability of 100% of electronic components?
Relying on AI and big data analytics, Cybord’s solution automatically inspects both the top and bottom of components and analyzes visual data and production metadata from the production line to ensure the authenticity, quality, reliability, and visual traceability of all electronic components.
The platform seamlessly connects to component assembly machines, allowing for near real-time inspection of 100% of the components in a fraction of a second before the assembly process on the boards is complete. By examining all components within the assembly process, Cybord’s solution empowers customers to identify and eliminate faulty components, which in turn reduces costs by minimizing the scrap and rework rates of assembled boards and providing smooth and efficient product delivery times. The software-based solution works with standard assembly machines, making it compatible with virtually every assembly line worldwide.
How do you foster innovation within your organization, and what role does research and development play in Cybord’s advancement?
The industrial market has become more dynamic, resulting in rapid fire changes that lead to tremendous successes when executed properly, but if they misfire, can lead to catastrophic outcomes. With assembly lines placing billions of components daily, companies must be innovative and efficient in providing the right solutions for this sector. Existing solutions and technologies are simply not enough to resolve the major pain points for industrial customers.
Therefore, to be a leader in the Industry 4.0 space, Cybord is constantly channeling the innovation that is the company’s beating heart in order to build products based on the most advanced technologies, like AI and big data. Because Cybord’s solutions are based on software only, our R&D team is consistently creating state-of-the-art designs and developing and implementing them effectively. From start to finish, R&D is our company’s main contributor to innovation and will always be our North Star.
How does your platform leverage AI to improve traceability standards in the manufacturing process of electronic components?
Many industry leaders appreciate AI’s capacity to improve various manufacturing processes, such as collecting data analytics, improving safety ratings, and enhancing security efforts. But many often overlook its potential impact on the traceability capabilities of individual electronic components. Today’s highest standard of traceability incorrectly assumes that all components within a reel are precisely the same, when in reality we have seen that approximately 90% of component failures are associated with visual indications on the components themselves. This is the critical gap in the market we are addressing. At Cybord, we identify those failures using the AI solutions outlined above, verifying that each component is procured from a reliable source, and that the homogeneity level within the same reel is high, free from any visual defects, oxidation, or mold evidence. By doing so, our platform is able to provide a complete inspection of 100% of the components that go onto a PCBA (Printed Circuit Board Assembly) and deliver the kind of surgical traceability needed in a zero-trust supply chain environment.
How do you ensure the proper documentation of electronic components, and how does this documentation help OEMs in their manufacturing processes?
The platform is installed onto a dedicated server within the Electronics Manufacturing Services (EMS), which collects all images from existing surface-mount technology machines (SMT) and automated optical inspection (AOI) machines. This innovation differentiates us from others as it requires no additional capital investments or modifications to production line layouts to deliver quality assurance and chip reliability. Cybord’s surgical traceability capability is based on AI inspection of visual evidence, which has been cross-examined with SMT traceability documentation in the Cybord AI cloud. This triggers alerts for any detected variations in date codes, lot numbers, or manufacturers. By establishing proper documentation in this way, Cybord is making this type of process an industry standard, with Tier 1 OEMs across industries relying on us for guidance in this area.
What challenges did you encounter while developing this visual AI platform, and how did you overcome them?
A key challenge we faced was the sheer volume and variety of components. Today, hundreds of millions of different component types are available for customer use, and trying to build reliable models for all these components is a nearly impossible mission.
As a result, we had to find a more innovative and efficient way to cover most of the components in use today. Our team is known for thinking outside of the box, and was able to resolve this challenge by learning to recognize the unique “fingerprint” of every component package in existence. This method makes the AI models smarter, faster, and significantly increases the possible coverage of these models.
What future advancements or developments do you envision for your visual AI platform in terms of ensuring even higher reliability and traceability standards for electronic components?
Cybord recently kicked off our real-time platform development. The next generation of the Cybord platform will be able to react within a few milliseconds and provide the component analysis results directly to the assembly machines.
Through this, Cybord will enable a real-time component ‘dump’ in the situation when a defective component is discovered. This strong capability will ensure customers are no longer required to re-work their products due to bad components assembled in these products. Preventing the assembly process of a bad component will increase the reliability of future products and reduce the risk exposure for every OEM adopting our technology.
Chief Executive Officer at Cybord
Oshri Cohen has over 20 years of vast management experience in operations, engineering, and supply chain in multinational high-tech companies. Before joining Cybord, Oshri headed supply chain at Nvidia Networks and, prior to that, held the role of global procurement and logistics at Mellanox (acquired by Nvidia). Oshri holds a BSc. in Industrial Engineering from Ruppin Academic Centre and MSc in System Engineering from Technion University.