Cybord, an Israeli AI software company founded in 2019 by Dr. Eyal Weiss, is positioning its technology as a fundamental shift in how electronics manufacturers manage component quality within the SMT assembly process. In an interview with EPP Europe, CEO Oshri Cohen outlines the origins of the company, the scope of the problem it addresses, and the challenges of convincing an industry that remains sceptical about the promise of 100% visual component inspection.
The founding story has a concrete and costly starting point. Weiss, who previously led a large-scale defence electronics project in Israel, encountered systematic field failures that were ultimately traced to a single defective capacitor valued at approximately one US cent. The incident highlighted a structural blind spot in PCB assembly: components are typically not individually inspected before placement, leaving manufacturers exposed to counterfeit parts, date-code mismatches, and latent defects that only surface during or after deployment.
Cohen joined Cybord in 2022, bringing experience from more than 15 years in supply chain roles at Mellanox and subsequently Nvidia following that company's acquisition. In those positions, he dealt directly with the consequences of defective components entering production — recalls, reworks, and associated financial and reputational damage. He describes the prospect of combining AI and big data analytics with a tangible supply chain problem as a primary motivation for taking on the CEO role.
Cybord's approach centres on integrating AI-based visual inspection directly into existing SMT pick-and-place equipment. The system analyses component images captured during the standard pick-and-place cycle, comparing them against a reference database to flag anomalies including counterfeit markings, physical damage, incorrect components, and traceability discrepancies. Because the inspection occurs inline and does not require a dedicated standalone system, Cybord presents the solution as non-disruptive to existing line configurations — a significant consideration for contract manufacturers and OEM assembly operations that run high-mix, high-volume programmes.
The claim of inspecting 100% of components in real time remains a point of friction with potential customers, according to Cohen. He acknowledges that many in the industry initially respond with scepticism, viewing the proposition as technically implausible given the throughput speeds of modern pick-and-place machines. Overcoming that credibility gap, he says, is one of the company's primary commercial challenges, even as installations continue to grow.
From a European electronics manufacturing perspective, the technology addresses concerns that have become increasingly prominent since supply chain disruptions of recent years brought counterfeit and substandard components into sharper focus. The automotive, industrial, and medical electronics segments — all significant in Europe — operate under strict traceability and quality mandates, making upstream component verification a compliance issue as much as a quality one. Cybord's software-based model, which can layer onto existing equipment from major pick-and-place vendors, aligns with the industry's preference for incremental digitalisation rather than wholesale line replacement.
Cohen's characterisation of the technology as 'sticky' reflects a broader pattern seen with inline data-capture tools in smart factory environments: once manufacturers have access to granular, component-level traceability data, removing that visibility becomes operationally untenable.
Quelle: Cybord – https://cybord.ai/news_events/news/epp-europe-interview-with-oshri-cohen-ceo-of-cybord-2/
