SemiWiki.com - Podcast EP313: How proteanTecs Optimizes Production Test

Episode Date: October 24, 2025

Daniel is joined by Alex Burlak is Vice President of Test & Analytics at proteanTecs. With combined expertise in production testing and data analytics of ICs and system products, Alex joined proteanTecs in October, 2018. Before joining the company, Alex held a Senior Director of Interconnect and Silicon Photonics Product… Read More

Hello, my name is Daniel Nenny, founder of Semiwiki, the Open Forum for Semiconductor Professionals. Welcome to the Semiconductor Insiders podcast series. My guest today is Alex Berlack, Vice President of Testing and Analytics at Proteontex. With combined expertise in production testing and data analytics of ICs and system products, Alex joined Proteontex in October of... of 2018. Before joining the company, Alex held a senior director of Interconnect and Silicon Photonics product engineering positions at Melanox. Welcome to the podcast, Alex. Hey, Daniel. Very happy to be here.

So you're an early employee. Why did you join Proteontex? Yeah, that's a great question. back then when considering this decision, part of my previous roles before joining Protanthex serving in different data analytics, product engineering, test engineering positions, it was always interesting for me how can we ensure that monitoring of the different chips and system kind of continues beyond the testing stages in order to ensure the quality and reliability and performance aspects to the customers. And actually, Proteontics mission statement was exactly about that, expanding this monitoring solutions beyond, say,

the traditional testing phases to in-mission. So that was very interesting and kind of personal angle for me, and I thought it can be very successful and wanted to be part of that. What do you see as the big challenges in system production testing today? So today, system production testing is much more complex. This is due to different aspects, considering the fact that there is usually a multi-chieplet environment in a package. The size of the chip is much bigger.

Therefore, also the assembly of the chip into the system and the interaction. between the chip and the system, having much more complexity around the different types of functional workload that is running, et cetera. And that actually requires a different thought process of system production testing beyond the things was existing up until the recent years. What is Proteontec strategy and application to address the market? it needs. So if you if you kind of look today of system production testing, most of the results and the tests are focusing on more of a pass-fail scenarios, meaning that you run the

system while the chip is of course assembled into that and you kind of get pass-or-fail results and decides where the system goes to the next step to shipment to the customer. For the Antex, actually thinks about that in a different way, in the sense that we provide visibility into what's happening in the chip and then subsequently also what is the interaction between the chip and the system. However, this visibility is done in a parametric manner, meaning that we integrate different types of what we call agents into the chip that is capable of measuring and providing different types of parametric signatures on critical metrics. that needs to be measured within the chip.

So there are two things that can be used using this type of data. One is a better understanding of the parametric both quality, but also some opportunities for optimizations in system level about the chip performance. So this notion of looking on the chip and the system from parametric perspective with using our deep data and the agents that are being integrated, is very different compared to this pass-fail approach that is used up until, you know, traditionally, I would say. What capabilities are you offering for system production testing? We have these two types of, I would say, two categories that we provide solutions or applications,

and it's separated or split around some software or applications that we provide. of the agents that have been integrated into the chip for making a real-time decision while the system is being tested. So this is one. And the second part is actually additional application that is able to analyze the results in a more aggregated manner across many systems. And this is a software solution that we provide in a form of cloud analytics platform that is able to aggregate a lot of data from different types of systems while they're running

in a production environment, to take even a higher level decisions and conclusions around the results and enable more efficient debugs, etc. So these are the two types, kind of in-line testing for real-time decisions per system. And the second part is overarching view of many systems and the ability to analyze all this data at one place in our analytics platform. And how is this different than current solutions? I mean, what are the benefits for the customers? So I think the main difference is this kind of parametric visibility enables much better control of both of the quality and reliability of the different systems and the decision-making around that. And the second thing is additional optimizations that can be done during the system.

production testing in the context of performance, power and performance of both the chip and the system. So if you kind of think about it in a traditional manner, today this pass-fail approach does not enable a very good debug in the case of failure, does not enable to detect systems that are marginal in this mode of operation when the chip is running with a functional workload, et cetera. So that can cause more failures later on in the actual mission of the system. And also does not enable, in the case of a failure, a proper debug and understanding where

the issue is coming from and kind of inducing a yield loss that is not fixable. So with ProtonTex approach, with the visibility that we provide throughout the different agents and the applications that are targeted to improve both quality and also enable better optimizations of the performance and the power during the system production testing. Now you can achieve a better capability to debug issues while the systems are running and in case of failures. But also it enables a better outcomes at the long term when the systems are operating in the field because we are able to catch different types of parametric signatures in terms of and prevent some quality escapes. And also,

eventually the performance and the power, so in the long-term in mission, it operates in a much more efficient manner. So I would say these are the major advantages, kind of utilizing this non-binary approach of past, fail, but actually looking on the parametric and the color of things during production testing. What markets are you targeting with the solution? Where do you see the most traction? We see a lot of traction in the data center segments in the context of AI chips and other types of data center chips. And we also see a big traction and utilization of our solutions in automotive because they care about quality aspects. In the data centers, it's both quality and performance. And also another segment is the telecom segment that also have interaction between the chip and the system.

And that's where also we see a very strong adoption of our solutions to address the things we discussed before. And last question, Alex. What trends are you seeing in the system market in general? Yeah. So I think we see a couple of things. We see that, you know, kind of hyper-scaler customers or hyper-scaler companies are developing their own kind of end-to-end solutions, including the chip and also having full control during the manufacturing and the design process from chip to the system and all the way to admission. So they would like to have a better control of the entire operation. So they're starting to become much more vertically integrated in that sense. And the way we operate, you know, it enables them to compare different types of parametric signatures from within the chip across the different stages in this vertical integration, starting from the chip level all the way to the system and then to, in mission. So this is a very interesting trend that is happening instead of more of like

segmentation between chip companies to system companies. We see in some cases some consolidation and vertical integration around that. So I think this is one of them. The other way is also happening that, you know, more traditional cheap companies are actually understand that they would like because of the complexities of the different types of workloads, they instead of only providing or shipping the chips, they are trying to move ahead and also manage and take responsibility on the system itself and basically becoming more of a system company. So this is kind of the other angle of that, that more traditional chip companies actually shipping and they're trying to get into the system world.

So that's another trend that we are observing. Great conversation, Alex. Thank you very much for your time. I hope to have you on again soon. Thank you very much, Daniel. I appreciate it. That concludes our podcast. Thank you all for listening and have a great day.