Hands-on Learning for PCB Design Students with Sid Deliwala

In this episode, we will talk about the importance of engaging and exposing the PCB Design students to today’s real-world industry demands. Our guest Sid Deliwala is the Director of Lab Programs in the Electrical and Systems Engineering Department at the University of Pennsylvania. Sid will share with us how he makes Electrical and Computer Engineering (ECE) fun and meaningful for our future innovators.

Listen to the Podcast:

Download this episode (right click and save)

Watch the video:

Show Highlights:

• Sid talks about his background and his advocacy for hands-on training 
• Tinker and play – learn the fundamentals, try new things and take risks
• Making the Electrical and Computer Engineering (ECE) major meaningful
  • Everything in the circuit world, in the hardware world, is also very data-driven
  • End to end experience for students, holistic understanding of Software, hardware, and PCB design
  • Preparing students to work in a multidisciplinary engineering environment
• If the cloud, the analog world, and all the digital circuitry that's needed to make things work rolled into one, it'll certainly be a very, very exciting time for all the future innovations
• You don't have to be an expert on everything, but you should at least have been exposed to everything in engineering
• Sid talks about the Penn Electric Racing – every year they participate in FSAE annual competition
• The problem with offshore PCB manufacturing – students have to wait for at least 3-weeks to receive their PCB for their project
• Importance of DFM and the difference it will make for students if they get the chance to visit the manufacturing company
• The supply chain and semiconductors shortage is real, the students experience current industry problems first-hand
• Sid expresses his sentiments toward a completely optimized system, it has some disadvantages

Links and Resources:

Sid Deliwala’s Director Profile, University of Pennsylvania
Podcast with Eric Bogatin: Pathological Design Features
Read Altium’s Newsletter: Staying on Top of the Supply Chain with Altium

Full OnTrack Podcast Library
Altium Website
Download your Altium Designer Free Trial
Learn More about Altium enterprise solutions
Altium 365: Where the World Designs Electronics

 

Claim the special offer for Podcast listeners only

Sid Deliwala:
The excitement for a student is not the board because they're also looking to put a phone there that can be downloaded from a cloud that has to be secure, I think they see the entire supply chain cycle.

Zach Peterson:
Hello everyone. And welcome to the on track podcast. We're very happy to have Sid Deliwala from university of Pennsylvania. He is the associated faculty and director of electrical and systems engineering lab programs at the university of Pennsylvania. We'll be talking all things, PCB design education today, and it should be a really exciting discussion. Sid Del, thank you so much for joining us. I'm very happy to have you here today.

Sid Deliwala:
Oh, thanks for having me, Zach.

Zach Peterson:
Absolutely. Absolutely. I think you've shown up on the podcast at a very opportune time, given the recent release of a report from Department of Homeland Security and Department of Commerce talking about the lack of workforce and the lack of overall manufacturing capacity in the US for PCBs and PCBAs. And so someone like yourself is on the front line of helping to deal with that.

Sid Deliwala:
Yeah. I cannot stress enough that... We lost quite a bit of time in our curriculum in last two decades, giving a priority to people who would be interested in manufacturing, partly because of reasons for cost and for labor, we promoted optimization in the revenue models. But I think if you make stuff, it's usually good for the economy.

Zach Peterson:
Sure, sure. I totally agree. Maybe before we get into your work at U Penn and the interactions that you have with students and manufacturing, maybe you can give us a little bit of your story and your background. How did you get into the electronic space and really what drives you to succeed in this area?

Sid Deliwala:
Sure. So my background is in computer engineering and I remember when I graduated from Drexel with a master's degree, I was a huge fan of electronics. I knew that software embedded back in late nineties, Zach, a few decades ago was very exciting. And I knew the word tech, but I never knew how important it would be in the field of academia. So I joined university of Pennsylvania with a vision to make sure that the programs and courses that support hands on learning would stay current. And that's what I do right now. My role is to integrate the curriculum, the hands on curriculum, and that means different things to different people. Traditionally we believed hands on with the ones where you use hands. And now that also means the ones that use keyboards.

Sid Deliwala:
So we are in a different world of hands on learning and the students currently treat that hands on education differently. So as you might expect that students who want to do electrical and computer engineering are also interested in the world of software and there is a huge need for people who can design, simulate and then fabricate. And that brings in a lot of software. And so I try to bring that excitement into the program in electrical and systems engineering. And I'm involved with two courses. One is a freshman class, which is for incoming class in electrical and systems engineering. And I also see students in their final year in the capstone for the entire year. So I have the opportunity to see students grow into the program and all their changes in the career that happens at career aspirations and to observe how innovation plays such a huge role in today's engineering programs.

Zach Peterson:
Yeah. You mentioned keeping things current. I'm sure that, as time goes on, that becomes a bit more difficult because things move so quickly. It just seems like... Back when I was in school, we were kind of focused on, MATLAB Mathematica, but now you have to know Python, you have to know Pandas, you have to know Matt Plotts, you have to at least know how machine learning works, things like this and build on those software skills even if you're a hardware person. I was in optics, which I think traditionally, you wouldn't think is something that requires a lot of programming until you start doing simulations, but moving into electronics, it's quite common to at least interface with a firmware person or a software developer and ensure that they know how the hardware works and you can properly explain it to them.

Sid Deliwala:
Yeah. But I also think Zach, sometimes we hold things back. The grownups are much more likely to hold on to things that they learned when they were in college. What I have seen is, when you expose, for example, freshmen to Google Colab and tell them that not only can you use sensors to interface to micro controllers, but you can write a code that changes the way you read the data. And I think that jump for understanding that a voltage is a data, is easier for the incoming student to understand them for us to appreciate. Because they see already data drives all of their daily tasks. So trying to keep things current, if we just level off with the expectations of our students and just look at how they use technology, we can eliminate the magic box of technology, why things work the way they are. The physics of it and the mathematics of it. They want to learn that. But you have to have that, excuse me, the maker spirit, which is let's do something first.

Sid Deliwala:
Whether it means software, hardware, doesn't matter, but that also includes simulation. That includes for them to tinker with Python, even if they don't understand everything about computer science. So I think we have to be opportunistic in letting them tinker and I'm sure it's risky. As instructor, I feel that there is going to be something uncertain about the course when we're things that may not work, but that's okay. If students see that there is excitement in the material, that's very current, they are willing to take that chance too, when they learn something new.

Zach Peterson:
Yeah. I agree with you. When it comes to allowing students to tinker and play with things and learn by failing or at least learn from failure. I was always the type of person that learned by taking a well documented example and reverse engineering it, and then using that to tinker. And I always felt that was a great way to get started and really dig into the guts of something.

Sid Deliwala:
Yeah. But we didn't have internet, Zach. The instructors always were the single point of contact for information. Yeah, there was library. But now, I remember in our last ESC111 class, which was the fall freshman class, we were going to introduce raspberry pie with a camera that would automatically detect color, and boy, it didn't work. Things broke, the hardware we got wasn't working right. And so we did the whole experiential side of it, but we never got a chance to do the lab experiment. However, in the final projects, students picked up things on their own. They went out, learned enough of pie and Python. Learned how to figure out how to use cameras. And we could support them in that adventure. But five projects out of 20 tended to be more about software just because they liked it. And I think they are willing to take risks. And sometimes we are the ones in the way of letting that take that chance.

Zach Peterson:
So what you do and when you refer to hands on labs and hands on instruction and hands on courses, things like this, I'm hearing that it's hardware focused, but you're really willing to have a broad inclusion or a broadly inclusive view of technology. It can include a software element. They may have to code something in firmware, then maybe they have to do something mechanical. You brought up cameras, perfect example.

Sid Deliwala:
Yeah. Yeah. I think it's about time. By training, I am electrical and computer engineer, but I think, ECE is a weird major. We try to teach everything from light bulb to how a computer works. And now how SOC is fabricated, there is no way we can hold on to the curriculum in which there is no software. I think it's a lot of time we do both. In fact, we can make the major more meaningful. If you tell them how everything they do in the circuit world, in the hardware world, is also very data driven. And I think in electrical and computer engineering, we are the experts on the interface. So we should be the one who should lead this data revolution. And of course, as you can see IOT, that's huge.

Sid Deliwala:
And I'm happy to have students entertain the idea of, edge IOT, edge intelligence and I see this shift. So students would go do an internship sometimes in Google, sometimes in Facebook or sometimes in very hardware companies. And they come back and they say, "Boy, I had a blast learning a deep learning network problem.". And these are EEs who have done circuits and they're like, my data is very valuable when it is in a cloud. And I want to learn that, it may make me a better analog engineer if I know how that data is going to be compressed. So it's a connection that students make very easy. We should let that naturally happen in our major.

Zach Peterson:
Yeah, I agree. I've often wondered and also talked with others about what it takes to get students more into the PCB side, instead of just focusing on, here's where you put components, here's how you route traces. Those are important tasks, but to really get students excited about hardware development takes a lot more. One thing I've been noticing lately is that some of the startups that I work with, they're actually being started by people from the software world, but they're realizing that they need to deploy this solution for the great idea that they have on a piece of hardware. And they don't even know where to start because they're software folks. And so I'm wondering, is software, especially with the crazy high salaries in Silicon Valley and things like this, one of the drivers that is really compelling for students to get into hardware design, do they start with software and then they realize, "Oh, this PCB design thing is pretty cool."

Sid Deliwala:
Yeah. I think if you remember, Zach, when I was in engineering, we made PCVs by dipping the board in a faid chloride solution or something. You handmade it. It didn't make our circuits better, but it certainly made it look better than connecting it with wires. So you felt like you made something that was wrapped enough to sell it to the world. I think the PCB bridges that connection, what students understand when they look at a PCB, I think is that there is a micro controller or a microprocessor sitting on that PCB. They don't necessarily see all the resistors and the capacitors that take care of parasitics, but the driving force for them to see why PCB matters is because they're excited about this. The world of SOC, the FPGAs, which are very software heavy. Both in programming them, writing formula for them, laying it out.

Sid Deliwala:
It's all very software centric. But if I wanted to sell it, if I wanted to make a computer, you will have to fabricate it with all the hardware on it. So even if we gave our students a taste of what it is to design for manufacturing it's just a huge step for them. They may not become PCB designers or they should. That's absolutely their choice, but they realize that, when you make a four layer PCB, it isn't trivial. All the design that they learned in their courses in circuits comes alive when they put a ground plane. And I think it's a great opportunity for them to build a product because they've always been building these prototypes. And now they can be proud holding that PCB and claim that they actually got it manufactured. So I think that's the end to end experience we should at least share with them. And sometimes it breaks my heart that I don't have enough PCB designers in United States who can do turnaround for students.

Sid Deliwala:
It's very expensive proposition, but I'm also hoping that at some point the insuring will happen and we'll have access to more local places for students to tinker and send their boards to.

Zach Peterson:
Yeah, definitely. And you brought up something quite interesting, which is, about that experience of getting all the way to the finish line and you have something that you designed yourself and then you power it up and there's that Eureka moment. Oh my gosh, it works. I always love that to this day. It does more than just turn on. It actually does what you wanted it to do. But what I'm hearing is that the experience of actually taking something that would've been like a dev board plus some modules and putting it all onto a custom board, let's say. That's an important educational experience and that's where the rubber really meets the road, regardless of whether or not they become professional designers. I think this is important because it prepares them to work in a multidisciplinary engineering environment that they may have to do, whether it's in academia as a researcher or whether it's an industry as a professional engineer. What do you think about that?

Sid Deliwala:
Yeah. Yeah. And the excitement for a student is not the board because they're also looking to put a phone there that can be downloaded from a cloud that has to be secure. I think they see the entire supply chain cycle. There is a cloud involved, even though the board that designed is a hardware and without the security in the cloud, that same day your 25 on your PCB may not be ready for the job. And so, you're really trying to connect all of these different pieces of design. And I think the biggest challenge in getting students up to the point where they're making a PCB is, it takes a lot of courses to get there. It takes a good circuits class. It takes a little bit of RF.

Sid Deliwala:
It takes a little knowledge of embedded. And then you also need to know a little bit of clouded AWS if you're storing your firmware on AWS. So I think that to me is, if possible we can deliver it at an undergraduate level or even a master's level, it'll be a huge step for our students to know that it's possible to do these things today and all from the computer and very little is needed when you're designing this than we use LTM, but there are other software that do the same thing.

Zach Peterson:
Yeah. I like what you're saying about the entire experience. It's more than just a board, because I think one of the things to this day that we've still got a combat with PCBs is the perception that it's just kind of an overblown way to wire up computer chips. And that's actually something I had to deal with. When I was teaching, most of students were electrical engineers or computer engineers. We had an ECE department at my university and none of them, nobody wanted to do PCB design that wasn't even a thought, like we have Intel here where I'm at in Portland. And so, everybody want to go put on the bunny suit and work in a clean room. I think that's what they thought they wanted to do. They probably changed later, but that's what they wanted to do.

Zach Peterson:
And of course the semiconductors are cool and the financial benefits are very compelling and you look at PCBs and I think it's kind of maybe a bit of a let down. But from the approach you're taking, where it's not just the board, it's the integration with the cloud, it's all of these other technologies that you can add to it. There's also a software component. I find that's actually a really compelling way to get students involved in the industry and hopefully rebuild or regain some of that knowledge and pass on to the next generation.

Sid Deliwala:
Yeah. And I'm sure Zach, you know that getting a high end board made is not trivial. We have a class called ESC five 16 taught by Eduardo Garcia. And the goal is to create a four layer single sided board where students will actually design the board, create a bomb, send it to the manufacturer, the whole process. And students who have done all the circus classes will certainly say, "Wait, is that what an antenna looks like on a PCB?". And if you forget to put it or forget to do it the right way, you're not going to get any wifi to work. I think it's also encouraging them to be better designers because they know the limits of putting a sensor on a board has more to do with just a sensor specification. I think if we can blend the cloud and the analog world and all the digital circuitry that's needed to make things work, it'll certainly be a very, very exciting time for all the future innovations that we have yet to see come through.

Zach Peterson:
Yeah, absolutely. I would agree. And my hope is that type of approach, whether it's by the CAD vendors or whether it's by the universities or even down at the high school level is really going to get more kids interested in electronics. Because I think one of the challenges that we've had, outsourcing and off globalization, not withstanding has been trying to get students interested in something that is not perceived to have the same minimum financial benefits as other areas of tech.

Sid Deliwala:
Yeah. And I think if you want to lead in 5G, you have to know what happens to high frequencies when you put them on a board. I think there are fundamental limitations to what engineers can design if they don't know hardware. I think it's about time we realize this and give it some importance.

Zach Peterson:
Well, what you're doing, I think on the front end, is actually very compelling because it's number one, getting students exposed to all of the stuff that goes into these advanced technologies, whether it's 5G embedded AI, maybe it's robotics, autonomous vehicles. We could go on down the list. So you're at least getting the students exposed to the mindset of how do we all collaborate to produce these innovations and then going further, I'm sure that students are going to gravitate towards one area of board design if they choose to go the route of packaging, or maybe they go into SOCs. Who knows what it is. But I'm hoping that, on the front end, you doing this kind of work and possibly other universities implementing similar programs, is going to get students exposed to the groundwork of actually working on that kind of team and being able to tackle those more advanced challenges.

Sid Deliwala:
Yeah. And what we also found was, we introduced a half a course in a class in our sophomore program, ESC292, where we just tell them to lay out a circuit board. It's the sophomore year. They don't know everything about circuits, but we bridged it with a CAD program. So that is solid works. So now they're exposed to making on a computer which they are gravitated towards. They don't want to machine apart, first. I think a lot of electrical and computer engineers like to visualize stuff. But it also opens up opportunities for them to say, "Well, I not only can design circuits, but I also can make objects." I think it's a design perspective. And sometimes bringing the element of design, it just changes a perspective that... In future, if I'm innovating for stakeholders, I'm innovating to make some new things. And you're not limited by your circuitry. You also have proficiency in some manufacturing.

Sid Deliwala:
It's a hard one to do because our curriculum is so packed. You want to teach DSP, you want to teach software systems, you want to teach embedded, you want to teach SOC and now to bring this skills is just a difficult thing to pack into a curriculum.

Zach Peterson:
Yeah. That's totally understandable. But I'm looking at this more on the end of what's someone going to do, if they maybe go into a graduate program. And are they going to have something like advanced packaging available as a route, or are they going to be prepared to at least join a team that is working on this advanced stuff and not have to come in as, it's either 10 years of experience or you don't get the job and be able to start participating and learning on the job and have all of the background that they need. Because when I was talking with Judy Warner actually before I started hosting the podcast, she had mentioned that she was talking with someone in the aerospace industry and what they look for, is you don't have to be an expert of everything, but you should at least have been exposed to everything in engineering. And so I like that the hands on route seems to address that kind of pain point.

Sid Deliwala:
Yeah. And we know that very well that devices might have reached their limits on sensing signal to noise ratio and you can have a front end, your sensor to be extremely good at signal to noise ratio. If it's not laid out right, you're not going to get the same output. Now, if I am in a team that understands the limitations of how the design would affect my front end, I think it affects the overall design of the product because I can do some of the trade offs that I didn't get in the PCB in the software. I can digitize before the PCB is able to alter the signal. So I think the idea that you ultimately are dependent upon, interconnects for some signal integrity is a great experience to have, whether you're doing SLC, whether you're doing FBGA design, I think ultimately it is limited by some interconnect that we bring in to the world of computing.

Zach Peterson:
Yeah. Yeah. I would agree. And you had mentioned in a sophomore class, just having students just go for it, lay out a board. This is interesting because we have another podcast that we did with Eric [Bogatin 00:26:31] and we'll actually link it in the show notes, but he had mentioned that he does the same thing. He just has students' layout, I think it's a little 5, 5, 5 timer. And they just do whatever they can to lay out the board and the way they think makes sense. And they can see the crosstalk and they see the lights blinking and all of that. And then they go back and they redo it. Do you have students go back and redo it with best layout practices?

Sid Deliwala:
We don't have that much time. They get the chance to build maybe one circuit board. By the time it comes back, they get to solder it. Remember that's another thing they have to learn because it's one thing to do through whole. And it's other thing to do surface mount. And when they look at this resistor that they can't find, when it came in a plastic bag, because it flew away when they sneezed on it. They certainly realized, boy, world has moved on. You may not be able to see all the parts that you see with your eyes, but that is something else act. And I'm really proud of it because these students who take this class, sometimes it's freshman also in spring. They go on to build other things on their own in student clubs. And I think we here at Penn were very surprised that there is a club in which they built an electric car.

Sid Deliwala:
It's called pan electric racing. They have to build a new car every year to participate in FSAE annual competition. So it's 50% EEs, 50% mechanical engineers, but it's an electric car. There is no way you could make the electric car without a PCB. This car has to steer. It has a high speed motor controller that maybe tens of amps controlling the motors. And there is a dashboard that has all the sensing going on. You're not going to have wires. So there is a team of students who are designing PCBs with HDMI output so they can seat on the monitor. And these are sophomores. So they have taken this to the next level without any, I would say adult interference, they picked up the software, they figured it out. Internet helped them learn it. Part of it may be our class, but they are on their own. And so who knew that the PCV experience will help them make a car on their own. And this is all innovation.

Sid Deliwala:
So they go race. They have to make sure the board survives all the integrity tests that happens in this competition. But I think they will take it to a point where they find it exciting. And you don't even have to teach that. Once you give them these skills, they're on their own to do this fun stuff.

Zach Peterson:
That's really incredible that students just a little bit of exposure to PCB and all of a sudden they're off to the races. No pun intended.

Sid Deliwala:
Yeah. What's more exciting? Remember, fast cars and rockets. Both are both are in.

Zach Peterson:
That's very cool. That's very cool. So I wonder if one of the pain points that maybe you have seen in delivering this education and then maybe getting students to take the jump into maybe doing their own types of projects, like electric cars or whatever it may be, is if it's a lack of resources or do you find that there is a good wealth of resources and maybe it's just a little difficult to parse through sometimes.

Sid Deliwala:
So there is definitely a pinpoint. Look at the world of software. If I wanted to learn deep learning on my own, I wanted to have a self learning class on Python, I have full access to space from AWS. I can do cloud computing and it doesn't cost me anything. Look at the hardware side of things. A decent board with four layers, four by four inch board to get manufactured with decent number of parts around $500.

Sid Deliwala:
And to give that experience, that's the only way of doing it. And I sometimes have hard time planning for this class. So Eduardo Garcia who teaches 516, plans as well ahead that the boards are ready now. They're being ready now, so they can send the board offshore, unfortunately takes three weeks, four weeks to get it and we'll have it in time before the end of the semester. What if I had access to this locally, I could get this boards maybe, or in a week. I would deal with someone locally. I would get the parts locally. I don't have to worry about sourcing all of these things across, worry about custom tariffs, but it is difficult. Plus it's very expensive. If I have 15 teams making boards, that's a significant financial burden to the university to support a class which is not typical that somebody would spend 25K on a class to teach people how to do PCB for the whole class. But I wish it was cheaper. I wish it was faster and I wish it was local. I hope you can get all three of them soon.

Zach Peterson:
Yeah. The local point you bring up that... I think that's valuable in two respects. Number one, students could just call up the manufacturer and learn a little bit of DFM through a little 10 or 15 minute conversation. That's always nice, you could have someone come into the class and actually deliver that. The other benefit is, I think we've said on the podcast few times, it's always great to go out and visit your manufacturer and actually understand and see what happens in one of these factories.

Sid Deliwala:
Yeah. I wish we had that... We don't do those things anymore. There are no visits in an engineering experience left. I remember many, many years ago you could go and see places getting manufactured, but I wish... Yeah, there is a video function and you can see it, but it's nothing like seeing a pick and place machine do it live and make this. Zach, I agree with you but...

Zach Peterson:
Yeah. I am the type anytime I get a chance to go do our factory tour, like where I was just at PCB west this last year and Summit Interconnect was nice enough to open up their doors to anybody that wanted to come and just get a tour. And I want to learn as much as I can about the manufacturing process. I feel someone who designs stuff, it's my duty to know, but it was really interesting to just see how they have their operations set up, learn about how everything moves through each department in the factory. And it was really interesting.

Sid Deliwala:
Yeah. And I think once a student gets a taste of what it is to integrate all their circuit design principles in a hardware, I don't think it stops there. I have seen students graduating from that class, doing better writing for form bear. They become experts in managing cloud and how they can do IOT. And then they're not afraid of laying custom PCB that could have a processor on it in which they could do some AI. I think the software world has figured out how to do resources, very, very well. And many of them are free. The hardware world struggles with giving things away for free. And I know it's not a easy switch, but I think that would be a great way to engage our students in engineering.

Zach Peterson:
Yeah. I definitely agree. Well, one thing that I hope you will be okay with us doing, is putting your LinkedIn contact information in the show notes so that if there are any manufacturers near where you're at or local, that would be willing to partner with you to work on this, I hope that they will reach out and find a way to make it economically feasible because I get exactly what you're saying. To produce a small board like that locally, it's expensive, especially if you want reasonable turn time. And then, having students go through the sourcing process is a great learning experience, but, oh, Lord, especially today with semiconductor shortages, I just had to deal with replacing a whole bunch of parts on a build and I wanted to throw my laptop out the window.

Sid Deliwala:
Yeah. Yeah. And you're right. Well, now they're seeing it. We know in our classes, we told them you can't use a certain part, even though it's in your item library, sorry. Those things are not in stock. And then they realize, boy, this is real. All the things they read about in supply chain and semiconductors is affecting my design, my board and my lead times. So it really engages them well. And I think we have just started this insuring excitement in the student community because they think we should make stuff and then if you make it right, I think there is market for it here.

Zach Peterson:
Yeah. I absolutely agree. What do you see is the future going forward to continue engaging students, especially as new technologies develop? Do you plan to try and find the easy way to integrate all of that stuff into your courses and hopefully draw more students in?

Sid Deliwala:
Yeah. I think we have had great success in engaging our students. What I find difficult is, Zach sometimes it's not enough to let them tinker with skills. I think we need to find a way to create some innovation model around all these skills. So if you tell them, "Okay, if you make this printed circuit board, you can manufacture in US, fine. Or you can think of it." Well, but are you going to create next generation camera here so that we can make a better sensor out of it? Are we going to make a better radar that can be better at detecting? And I think if we tie in what they see as the future, it does better job of representing the innovation that is waiting to happen. And I think things are just beginning in the world of AI.

Sid Deliwala:
There is this whole edge AI that's happening, where everything's integrated at the edge, all things go in the cloud, but then everything is custom. SOCs are custom. And we have to take them through that step and just let that happen in a different way to... They should probably know how to do some layouts. They should know how to do PCB. They should know how to do high frequencies, circuit design. And if you show them all these different ways to design, it just makes them better and gives them more opportunities to do something cool in their careers.

Zach Peterson:
Yeah, absolutely. And then I think the one thing I'm wondering, especially given the report that I mentioned at the beginning of the podcast is what's barring students from getting into manufacturing. What's the big hold up? Is it really just that the economic model is so screwed up and we need and companies have to properly price in the risk of going overseas? Or is it something else? People don't want to work in manufacturing. So of course they have to produce overseas because they can't find anyone local to do it. You can't produce locally. Is that also part of the problem or is it just everything?

Sid Deliwala:
I think they over optimized everything. Last 20 years, we kept just enough inventory so we use it. We just kept enough resources so we can order it when we want it, not realizing that even the slightest disruption can throw off an over optimized system completely. So that's engineer's fault, right? We try to calculate the model for the best optimization for everything. Including costs and revenue, not realizing that there are huge advantages to not having an optimized system and having some resilience in the system. I think after COVID, people appreciate the resilience that is a semiconductor shortage that's going to last into maybe 2022, 2023. So maybe we will optimize less. Maybe it'll be okay for us to be not perfect in manufacturing and all the cost savings. So I hope that happens. It's not just about independence. It's about having resilience in the system so that it doesn't come to a grinding halt. Like you can't go and buy a car right now, because there are no cars. That is a result of that.

Zach Peterson:
Yeah. That's a problem. And I guess the natural result of that, is by having a slightly un-optimized system, or maybe by broadening the set of considerations beyond just driving cost to the bottom of the barrel, you now are required to produce in places where you now have to have people local, because at... It's true you don't have 100 people running an SMT line, but someone has to run the line. Someone has to do the inspection. Someone has to do the front end cam, somebody has to manage everything. So people are still involved. So I guess that's the natural consequence and it creates that demand.

Sid Deliwala:
Yeah. And maybe it'll be a time and there may not be enough people to do all the work. There is no doubt and computers will probably do much of it with automation, but at some point we'll decide as engineers how much to do that. It may be a rational choice to keep people doing things. But I think there lots of opportunities. Things like health tech, where there is a long way to go before innovation appears to be overdone. Like, do we really need a phone that goes at 50 pictures per second in terms of frame radar. Do we really need to over-optimize and overdevelop things that are already at a point, which is really, really good. I want to make health devices that does better diagnostics. But I think at some point when we have this freedom of deciding where to put our resources, well, I'm sure we'll figure it out like we've done it.

Zach Peterson:
That's all great points. Well, thank you so much for joining us and thank you for the work that you do. I couldn't stay in academia forever. And so I definitely applaud the people who do and who are working so hard to educate the next generation of engineers. So thank you very much for everything that you doing.

Sid Deliwala:
Oh, thank you, thank you. Nice chatting with you. But the secret here being in academia is you're surrounded by fountain of youth. So you missed out. Okay.

Zach Peterson:
That's fair. That's fair. Well, I'm hoping one day I can make my way back into academia truthfully. So we'll see what happens. Thank you so much for joining us and to everybody in the audience, please check out the show notes we'll have some great resources there where you can connect with Sadar and learn more about what they do at UPenn. And we'll have some other great resources that talk about some of the other, or that point to some of the other things we mentioned in the podcast. Thank you again. And to everybody out there listening, don't stop learning and stay on track.