Ep 42: Enlightening the Future of AI and Photonics with Excelitas Technologies EVP & CIO Amit Shah

‍Evan: Hi there, and welcome to Enterprise Software Innovators, a show where top tech executives share how they innovate at scale. In each episode, enterprise CIOs share how they've applied exciting new technologies, and what they've learned along the way. I'm Evan Reiser, the CEO and founder of Abnormal Security.

‍Saam: I'm Saam Motamedi, a general partner at Greylock Partners.

 Evan: Today on the show, we’re bringing you a conversation with Amit Shah, Executive Vice President and Chief Information Officer at Excelitas

Excelitas is a global technology company specializing in generating and sensing light in a field called photonics. They build optoelectronic components that power a range of products from home security systems like your Ring doorbell to military aerospace technologies (like a HUD in a fighter jet)

In this conversation, Amit shares his thoughts on innovative applications of AI at Excelitas, the impact of Generative AI on productivity and cybersecurity, and the future of enterprise AI.

Thank you so much for joining us today. I know we've had a bunch of chats in the past and excited we get to, uh, talk and, uh, share some of your experience and wisdom with others. Maybe just to kick it off, do you mind giving our audience a bit of background about you and your career and how you got to where you are today?

Amit: Sure. So I've been working for Excelitas, as CIO since 2015. Before that, I was with Excelitas as VP of IT and, before that, director of IT. Excelitas was spun off from a company called Perkin Elmer. Perkin Elmer is, or was, a component of S& P 500. So, we have a foundation in a very well established, what I would call life sciences technology company.

My background is actually in engineering. My very first job was actually as a test engineer working for a semiconductor equipment manufacturing company in the Bay Area. And that company acquired another company, and they, again, they were going through integration of their IT system. This is in late 90s, and they were running out of programmers.

If you remember, we had a thing called Y2K back then. And I always wonder how corporations work. And obviously, as an engineer, I had a affinity towards coding, programming, and just understanding how to use technology to create products or processes and outcomes. 

I love the idea of using business technology to create business outcomes. Technology on its own is great, but in the end, for me and people in my position, it's all about outcomes.

Evan: I wonder if you could share a little bit about, um, Excelitas.

Um, you know, you guys have a pretty wide range of products that impact probably more people than they realize. For maybe some people in our audience that are less familiar with what you guys do, do you mind sharing a little, you know, a bit of an overview of the kind of business and some of the product and services you guys, uh, deliver?

Amit: Absolutely. So, at a very high level, Excelitas is a company which operates in what I would call photonics and optoelectronic space. So we have a global network of design and manufacturing locations in America, Europe, and Asia, right?

We have a, like, a diverse portfolio of photonic technologies and products like lasers and light sources. Optics and optronics and imaging products. We also have sensors and detectors and we also do some electronics like power supplies and what we call ballistic electronics. Electronics which goes into ballistic systems like missiles and so on and so forth.

A lot of our customers are some of the biggest names you know in the industries we serve. So we serve on average I would say five to six different industries. Aerospace and defense, industrial, medical, Safety and security, semiconductor equipment manufacturing, and smart building technologies as well.

Right? So, and when you think about what we make, so we make products which creates light. We call ourselves the light people. So when I say light. Doesn't necessarily mean the visible part of the electromagnetic spectrum. I'm talking about UV, deep UV, IR, deep IR. So essentially we play in that area. So how to create IR as a light source.

How to create UV as a light source. How to bend that light. So the bending part comes through lens and what we call opto mechanical assemblies we make, right? So you have to do a lot with light to make it work for your particular application. And then we make sensors which sense this light. So you can sense UV and IR using semiconductors as well.

So a couple of example for your audience to relate to. A lot of home security platforms, Ring and Google Nest and so on and so forth, they work on principle of essentially measuring or identifying an object versus a living thing, which has a particular heat signature. And you do that using IR sensors, so we make those kind of sensors.

So when you walk past somebody's house and it starts recording, it's essentially measuring, for lack of a better term, the IR profile of that human being and telling an ASIC, okay, go start recording. That ASIC tells the sensor, start the camera, start recording, upload to in the cloud. All this happens in a matter of milliseconds, right? So we make sensors which goes into those kind of applications. 

On the other hand, when you think about optics, you know, some of the Most rugged optics, which we make, goes into planes. We make heads up display for commercial and fighter jets. So, when you see, you know, Tom Cruise flying that F 16 in one of those Mission Impossible movie, and you see that green thing in front of you, Well that's called HUDs, heads up display.

We make a lot of optics which goes into making those HUDs. We also make night vision goggles which you might see in a movie where a soldier on the battlefield is using to identify objects. So we make those kind of night vision goggles and we make light sources which goes into endoscopy and some other medical applications.

So when you think about Excelitas, right. We deal with light. And especially UV and IR. While we might not have a product with our logo, we actually have a few, but for most part, we are what we call subsystem and component manufacturing company. But we are again IP driven technology company. We have 1400 or so patents in all of these technologies and our technology is used in a lot of product we use as a consumer and businesses and medical facilities. And it kind of empowers or enables those products to do what it needs to be done. 

Evan: And Amit, you and I would be both remiss in our jobs if we were on a podcast in 2024 and didn't talk about AI.

AI is kind of tricky because it does feel like there's a ton of hype and like everyone's AI company in 2024. But at the same time, it does feel like there's kind of real world value delivery that's happening because of the technology. Love to hear your thoughts on where you feel like we are in that hype curve and if there's any examples where you feel like you guys have, you know, how to material win in terms of like developing products or operating your business with AI and maybe there's other areas where you're kinda a little more bearish on that you could share with our audience.

Amit: So I think a lot of hype recently is around Gen AI, as you and I know right? All this LLM models since open AI came with their LLM back in I think 2023 March, if I remember correctly. You know, the hype cycle is, is in full gear when it comes to Gen AI. But I, I take a, a broader view when I use the word. AI, right? 

I think at the very lowest level. I would argue RPA, and we use a couple of RPA platforms. That's AI, and we've been using RPA for almost four years now. We have significant automation of some of our core business processes, which connects to multiple enterprise software and applications. Gives productivity to, to our data analysts, business analysts, people in finance, people in HR, people in marketing and IT. So obviously we have been doing that for, for a while.

We've been using fairly complex and very sophisticated machine learning algorithms to make some of our products better. So we are actually going to launch, hopefully soon, some of our NVG, night vision goggles, which will be infused with very complex machine learning algorithms, which will aid a soldier on the battlefield in identifying objects. So that's one example. Another example, we are consuming those kind of AI driven or machine learning driven optical inspection machines.

So when you want inspect products we make, you know, we use machines which are powered with what we call machine vision technology. And it is using machine learning algorithms to help improve your QA inspector's job. So rather than doing it all manually, now you're aided by some of this machine learning applications as well.

And very recently we are evaluating the Gen AI for our employee populations through a bunch of co pilots, you know, built by some of the biggest enterprise software companies in the world, but also for our developer community. We have, you know, close to 100 developers, and we believe there is a provable value creation opportunity to enhance the productivity of folks who deal with the code.

So we are working in that area. We believe that Gen AI will have its moment in enterprise, but to get there, you need to have some foundation done right. And I think nothing is more important than getting excited, getting your people excited to, to adopt some of this technology and getting your data and processes cleaned up so you can actually get better outcome, right?

There's a golden principle in IT world - Garbage in, garbage out. No matter how good the technology is, if you're feeding garbage to that technology, it's just going to spit out garbage. And AI will, as we know, hallucination is the word people use. And when you think about what is hallucination, and what it means to an individual in an enterprise?

To me, it boils down to the input is not good enough, hence, you're going to have all this output, which is not good enough. So, you know, cleaning up your data sources, making sure your processes are streamlined, well understood before you adopt Gen AI is going to be very crucial. But having said all this, right, AI is a value creating, in my opinion, as Major shift as we had back in late 90s when we all started using internet to run our businesses.

So this is going to shift the productivity curve significantly. We're going to see staff function improvement initially, maybe slowly, but then I think it's a nonlinear function. Your productivity is going to grow exponentially. And if you don't invest now, it will be a competitive disadvantage. I don't know if it's a competitive advantage to us right now, but I strongly believe if we don't invest now and build foundation to take advantage of AI, we'll be at a significant competitive disadvantage in next half a decade or so.

Evan: When most people think about AI, you know, they do kind of focus on generative AI right now. And even generative AI, I think about text generation, picture generation, video generation, right?

But I think one reductionist way you could think about AI, it's by taking some data set, it's translating, converting into a different data set. The data set could be, You know, a judgment or decision or some autonomous actions. And so maybe in your, you know, again, I don't know much about the technology, but presumably with night vision goggles, you're kind of reading in data and kind of a different spectrum of light. And then you're converting that into some representation that's helpful for the user, right? Probably same thing in heads up displays. 

So I imagine like that, that kind of concept of using algorithms and technologies to kind of simplify translate to make the end user more productive, that's like a thing that you guys have been doing for a long time. So do you see kind of AI is just accelerating some of your technology across these products to do just a better version of that going forward? 

Amit: The simple answer is absolutely yes. And the way we think about this is, you know, I put this in two different macro categories. One, I call analytical AI, which is your classic machine learning, RPA stuff, right? Well defined, very narrowly defined problem and very well proven tool set to achieve the outcomes you need to achieve. 

And then this whole huge productivity machine, Gen AI. Which is still in nascent stage, in my opinion, and Again, the inputs are not clearly defined. You still have to work with, you know, huge data set to create outcome, which a lot of time I question, but I'm sure it's going to get better. 

Majority of our focus and realized value has happened in what I call analytical AI framework, right? So you talked about night vision goggles, you know, reading sensor data, improving your predictability. So when you are a soldier in the field, right? When you are looking in pitch dark. You're looking at a tank or a machine gun or, or some kind of military object. You want to have like certain degree of probability and say, okay, this is 99 percent probability that this is an enemy tank or friendly tank.

Right now, that disposition is done by soldier without any aiding coming from technology. Now we are going to aid that soldier with technology and improve the probability because we're talking about life and death situation here, right? So every percentage improvement in the probability matters a lot to, to, to the governments who pay for our product.

So absolutely, AI has created step function improvement and a proven value, which people are willing to pay with their dollars. And I think that's the most important thing, right? Technology for sake of technology doesn't mean anything. It has to create outcomes. It has to create value for the customer. We are very value oriented organization. 

Evan: So, I want to kind of ask you to pull out the crystal ball for a second, right? Earlier you mentioned the big trends being both AI and cybersecurity.

Certainly five years from now, both those dynamics and environments are going to be radically different than they are today. When you think about kind of the development of these new technologies, like how do you see that affecting the world five years down the road? Right. And how do you see that effect affecting kind of running, you know, running your business, but also building, you know, some of your products and services? And, you know, what opportunity do you think AI has to dramatically improve that?

Amit: I would say three things. First and foremost, the productivity equation is very clear to a lot of CIOs where using Gen AI and, and all the other AI types, for lack of a better term, is going to improve the end user productivity, right? Within a company within an organization. 

And when you think from a macroeconomic perspective, I remember Alan Greenspan back in nineties talked about how IT and IT and communication technologies, right? We're talking about mid 90s when we're just inventing routers and switches and connecting different networks with each other, to this thing called internet and how that improved productivity for everybody.

And today, after 25 Plus years, we do everything on the internet. This podcast we are doing is recorded on internet. Commerce is done on internet. So many things we do on our daily lives, in our daily lives is done on internet, right? We're talking about music, we're talking about watching entertainment, content, content generation, content distribution, content consumption, all of those things are done through internet.

The reason I'm bringing this example, I think AI is going to be like that. It's going to be infused. You will be using AI without you knowing you are actually using AI within a corporation, right? It will be embedded into your data structure where it might be cleaning the data or advising you how to improve your data based on independent analysis done using some machine learning algorithms.

It will help you execute your business processes faster, close your month faster. Being able to distribute your documents and other data to your employees, and your customers, and your vendors, and your partners, in a much more secure and efficient manner. I think all those things will be very foundational in highly, highly adopted throughout the world in the next five or so years. That's, that's my prediction. So that's one. 

Two, I think cyber is one area where it would be shameful if we don't use AI to make cyber outcomes better. Even today, you know, cybersecurity is one of the most complex thing a CIO does. I have to deal with 20 different vendors, and I'll be honest with you, I want better outcomes from my cybersecurity vendors.

I want them to use AI because, let's just be honest, the bad guys are going to or already are using AI to be more effective in their outcomes. I want my cybersecurity vendors to make sure they are using AI to reduce the complexity in the fragmented nature of how cyber security is done today. I think we have all reacted to the cyber threat rather than strategically thinking about how to create cyber security solution, which helps CIOs and CISOs protect their information assets.

Nobody can read billions of logs. I'm sorry. SIM doesn't work. I'm sorry. It doesn't work. Yes, we have one, too, but it doesn't work. And by the time I react, somebody is already in my network and doing things I don't want him to do. I want AI to spot that in a millisecond and prevent that. Right. So we need AI to be used in cybersecurity solutions. I think that opportunity in itself could be worth hundreds of billions of dollars. I strongly believe that. 

Number three. I think programming is going to be totally different going forward, right? I remember the days when I used to write C code and the kind of things I used to worry about, things have changed significantly. Now, my daughter, who knows Python programming, creates better code than I created when I was doing C code. And that's without any aid coming from AI. 

I think this is an area where creation of code is going to be revolutionized. And that means the innovation is going to be even better and faster because now you can create, test and deploy the code much, much faster, right? So your innovation cycle is going to be shortened for software and everything runs on software. So I think That's going to be an amazing opportunity to seize for everybody, right?

In the end, if you're using any IT technology, you're a software company, right? Yes, we make physical products. Some of our physical product goes with firmware or software. Some of our physical product doesn't. But every product we make, is touching some software. So in the end, we are a software company and we have developers and we need to make sure those guys are protected and have access to this amazing AI enabled programming coding technologies, right? So I think this is, that's another area where I think you're going to see significant improvement and a proven value creation. 

I think in the end, I always talk about technology being tool. What we are really supposed to do is create good outcomes for our employees and our customers. And AI will kind of give us better tools to create better, cheaper, faster outcomes for our employees and our customers. That's how I feel about AI in the next decade or so. 

Evan: So I think it begs the question then, both for you as a CIO, but for a lot of your peers, how do you best prepare for that future? Right. How do you equip the workforce? How do you build the infrastructure? Right. I, I just love to hear how you see that changing.

Amit: What we've been thinking about internally and what we've been presenting to even our board is we want to simplify and harmonize our data silos, right? So what I mean by that is, you know, we are acquisition driven company. So we have a lot of data silos. We have three different CRM platforms, 11 different ERP platforms, 20 different cyber security solutions. We are making a concerted effort on simplifying that. 

Your data silos needs to be as fewer as possible, because I believe for you to have better AI outcomes, you need to have good data going into that system, especially for Gen AI. So, reducing the complexity of our data silo is number one. 

We are a decent sized company, 9000 plus employees, 25 locations, 12 different countries, so we need to come up with things which are federated, but complies with different cyber and privacy regulation, right? So one of the biggest concern I have today, in terms of adoption of Gen AI would be privacy and security.

So I want my cyber security vendor to give me assurance that if I go out and, you know, run my workloads with one of the hyperscaler for my AI needs. It's going to be secure. Number one. Number two, it complies with GDPR and other regulatory, CCPA and so on. And every country has their own framework. Every state in U. S. has their own framework now. So how do we comply with those privacy and security regulatory framework is going to be very important.

Being a defense company, if you have a breach, now, if you're a public company, you have to report within 72 hours. Or 48 hours. In my case, I need to DOD and inform them what happened. If we had a security incident, we have to follow very well defined and very strict protocol. And for us, that's very important because if, if you don't protect our assets, nothing matters.

So we need to, we as IT organizations, CIOs need to create good, well defined guardrails, but then empower different teams, different organizations within your company to take advantage and have them articulate the value creation opportunities they might have for what they do using this tool. 

Evan: Okay, Amit, I have like a hundred more questions for you. I got through like one fifth of what I wrote down ahead of time, but we are a little bit short on time. So I wanted to kind of jump into a lightning round. So looking for kind of your like one tweet responses, just so we get some kind of, you know, punchier, punchier answers. 

Amit: Sure. 

Evan: Alright. So first, first, first lightning round question. How should companies measure the success of a CIO? 

Amit: Three things, top line and bottom line value creation. Very important for shareholders. So that's number one. Number two, cyber is a big part of it. If you can't secure information assets, you shouldn't have that job, right? So I think that's number two. Number three, employee enablement so they can create productivity in their day to day jobs. 

Evan: What's one piece of advice you wish someone told you when you first became CIO? 

Amit: Don't get bogged down by every little thing. Focus on big picture and be ready to be corrected. and be ready to change your direction.

Evan: So, one reason I was excited to have you on the show is you have a lot of experience, but also you've had a lot of experience kind of in one organization that is very complex. I imagine you wouldn't have been able to have the same success you've personally had or the company has had, unless you kind of had great relationships and great collaboration with your peers across the executive table. How should CIOs best position themselves to collaborate with the rest of the C suite?

Amit: CIOs should consider themselves as enabler, first and foremost, your job is to enable. different organizations to figure out how you can enable different organizations and their leaders to do their job better, cheaper, faster. So I think that would be an overarching answer to that question. 

And also think about how technology driven solution can either augment or create your existing business models. And we had some success with IT or technology driven business models. And once you have a successful case of that, you will open a lot of eyes. And for us, you know, we are in what we call industrial technology sector, which may not be as modern when it comes to some of the other sectors like software, enterprise software, or, you know, Silicon Valley style companies. 

Once they see the real value creation using I. T. And when I say real value creation, I'm talking about top line and bottom line. You will see an amazing demand for your talents all of a sudden. So think about yourself as an enabler, first and foremost, and then like an entrepreneur within your own company and say, look, you know, I'm going to do something which we have never done. I might fail, but if I'm successful, all of a sudden there'll be demand for my services and the things which will enable us as an organization to create value for all stakeholders. 

I think the entrepreneurism within a structured corporation like ours is very important. Being entrepreneur doesn't mean you have to be an investor driven company. You can be an entrepreneur within a multi billion dollar company too. And, and I think it's a mindset more than anything else, to be frank with you, and people who have that mindset, they tend to be more innovative by nature, right?

Willing to try and fail, and learn from their mistakes and move on. And I think that kind of growth mindset is very critical to be a successful entrepreneur within a corporate structure. 

Evan: So I want to switch a bit more to the personal side. Um, is there a book you've read that's had a big impact on you, and if so, I'd love to hear why. 

Amit: One book, and I'm going to give you a business book, which has still significant impact on how I think, and it's classic. Everybody has read, read that book, and I think everybody should read that book. It's Good to Great by Jim Collins.

I think it talks about how you can take a good company, make it, Great enterprise and the things which separates, you know, good company from great companies. There are a lot of great books out there, but even today, for me, that book gives me amazing view of business world, which I appreciate a lot. 

Evan: Sticking on the personal theme, what is an upcoming technology that you're personally most excited about?

Amit: Robotics. 

In the end, AI running on its own will help with our white collar jobs, but if you infuse that AI, which can create Kinetic output in a much more intelligent way. Boy, oh boy, you're going to live in a very different world. 

And in our company, we actually use robotics, not infused with AI just yet. But I was thinking, what if you can have an intelligent robot? It doesn't have to be humanoid. It could be any pick and place robot, right? But all of a sudden it's infused with AI. You're going to have another step function improvement in anything you do. 

So I think RPA, which is robotics process automation, and I see the value. Now I can just take that and say, okay, what if I can do it in a warehouse? What if I can do it on shop floor? What if I can do it in space? What if you can, if you can send humanoid robots to Mars before you actually send humans and they do all the functions human will eventually do? I think you're opening up a whole different possibility of doing things which you can't even dream of today.

I think robotics is going to be very different because of AI in next decade or so. 

Something I've been studying lately, Silicon Photonics. It's an amazing field and getting a lot of VC funding lately. And we play in, in that area. Quantum cryptography is another area, right? The whole area of cryptography is going to be completely different in the next 10 years, people have no idea. 1024 is nothing anymore. You're going to have encryption levels which, which is going to be so, so good. And, and I think that's going to be required because the world we're going to live in, where AI is used by criminals, we need to be ready. We need to have better encryption than, than the criminals. So I think you're going to see significant improvement in that area as well, quantum cryptography. 

Evan: I do know that, you know, once we have all these amazing robotics, they're going to need some extremely sophisticated, you know, optical sensors. So, uh, hopefully, uh, you guys will be ready to help us with that. 

Amit: We are working on it. 

Evan: Well Amit, thank you so much for joining us today. As always, it's been a pleasure to speak with you and let's find a time to chat again soon. 

Amit: Thanks for having me.

Evan: That was Amit Shah, Executive Vice President and Chief Information Officer at Excelitas

‍Saam: Thanks for listening to the Enterprise Software Innovators podcast. I’m Saam Motamedi, a general partner at Greylock Partners.

‍Evan: And I’m Evan Reiser, the CEO and founder of Abnormal Security. Please be sure to subscribe, so you never miss an episode. You can find more great lessons from technology leaders and other enterprise software experts at enterprisesoftware.blog.

‍Saam: This show is produced by Luke Reiser and Josh Meer.

See you next time!