EMS@C-LEVEL · 31 Jan 2023 · 7m
Cogiscan CEO Vincent Dubois discusses how recent supply chain disruptions have forced a cultural shift toward open collaboration between EMS providers, OEMs, and technology partners. He highlights the transition from transactional vendor relationships to long-term partnerships focused on solving specific manufacturing issues, such as reducing false calls on AOI machines. Dubois emphasizes that successful AI implementation in electronics manufacturing requires combining software capabilities with deep domain knowledge of SMT factory floors.
Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.
Manufacturing
Cogiscan applied AI and machine learning to distinguish false calls from real calls on AOI machines, reducing the need for human intervention.
“the use case I was talking about, for example, that we did last year is on AOI machines where you have a lot of false calls. So how do we sort a false call from a real call? And how can you reduce the human intervention needed to do that…”
Speaker 2 · Hear it at 5:40
early-stage companies
Other shows on the same ground, closest first.
Bryan Cantrill and Adam Leventhal of Oxide Computer Company discuss scaling manufacturing with colleagues CJ Mendes, Kirstin Neira, Erik Anderson, Aaron Hartwig, and Doug Wibben. They detail the transition from low-rate NPI to volume production using contract manufacturer Benchmark Electronics, highlighting specific engineering solutions like the Medusa and Reverso test boards to address fragile high-speed cables and press-fit DDR sockets. The team emphasizes their strategic procurement approach, building deep relationships with suppliers like Semtech and Texas Instruments to secure supply during the AI-driven chip shortage, while navigating compliance and tariff challenges for global expansion.
Kate Hicks, Head of Operations at Oxide Computer Company, details how her team navigated the global supply chain crisis by integrating operations early into the engineering process. She explains specific strategies for securing long-lead components like Sanyo Denki fans and NXP LPC55 microcontrollers through deep vendor partnerships, broker market vetting, and custom modifications. The discussion highlights the importance of API-driven automation using tools like Duro and OrCAD to manage inventory and forecasting for a small, scrappy team scaling production.
Oscar Lovera, COO at Xometry, explains how the AI-powered manufacturing marketplace is digitizing the sourcing of custom parts by integrating design, quoting, and manufacturability feedback into a single workflow. He discusses the shift from traditional, disconnected RFQ processes to instant, scalable matching between engineers and a global network of manufacturers, including CNC, 3D printing, and injection molding partners. Lovera emphasizes that while AI automates repetitive tasks, human expertise remains critical for complex problem-solving, trust-building, and ensuring operational excellence, arguing that strong processes matter more than flashy equipment. He also shares his perspective on supply chain resilience, advocating for global flexibility over local sourcing and highlighting the importance of intentional career choices for future engineers.
Russ Bukowski, President of Mastercam, explains how AI-powered CAM systems like Mastercam Copilot are bridging the knowledge gap left by retiring manufacturing experts, compressing the learning curve from years to months. He discusses how mid-size machine shops are gaining leverage through vertical integration and exclusive supplier agreements, driven by post-COVID supply chain vulnerabilities. Bukowski also details Mastercam's acquisition by Sandvik, which enabled significant investment in new technologies like Everpath and the acquisition of channel partners to foster direct customer relationships.
Phil Marcoux, Advisor to the Printed Circuit Engineering Association (PCEA), discusses the transformative role of AI in the electronics industry, emphasizing that successful AI integration relies on high-quality, human-generated data and closed-loop feedback between manufacturing and design. He warns against the current 'AI bubble' where startups receive funding without defined products, drawing parallels to the dot-com era and the early adoption of Surface Mount Technology (SMT). Marcoux highlights specific applications such as component availability checks via tools like Wenovo and Circuit Mind, and the critical need for standardized communication protocols like IPC CFX to bridge the gap between design and production.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by EMS@C-LEVEL. The episode itself belongs to its publisher.