Episode 9
Humans, Robots, and the New Era of Innovation
With Kevin Derman
What if doing more with less is the real AI edge?
In this episode of Rewired, host Akanksha Bilani sits down with Kevin Derman, National Strategic Alliance Director for AWS at Slalom and head of their AWS AI Innovation Lab, to explore how Slalom is helping customers move from GenAI ideas to real-world impact. From reimagining customer experiences in the airline industry to transforming how leaders think about cost, performance, and silicon, Kevin’s work shows what’s possible when innovation meets intentionality.
Together they uncover some of the most persistent myths around CPUs vs GPUs in AI workloads and take a look into a future where robots navigate the physical world, humans are augmented beyond devices, and intelligence becomes woven into everyday life.
Because when AI evolves, so does everything around it.
REWIRED Ep 9 Podcast Transcript
Akanksha Bilani (Host):
Welcome to Rewired! A podcast that shares the coolest stories of innovation and behind the scenes breakthroughs from Intel, AWS, and our favorite partners. I'm Akanksha Bilani, your host, and I run the go-to market team for Team Amazon at Intel, helping bridge the path with AWS partners and customers. Today, I have the privilege of speaking to Kevin Derman, the National Strategic Alliance Director of AWS at Slalom.
Kevin, thank you so much for being here.
Kevin Derman:
Hi, Akanksha. Great to join you today.
Akanksha Bilani:
I would love for you to tell our audience a little bit about your role within Slalom.
Kevin Derman:
My role at Slalom is to manage the partnership that we have with AWS, Amazon Web Services. Also fantastic partners like Intel, yourself. And ensure that our go-to market strategy is aligned and that we're doing things together that are benefiting our mutual clients. The other role that I have there is I head up our Slalom AWS AI Innovation Lab, which demonstrates these emerging technologies and new technologies in the AI space for our clients and really makes it real for them.
And if I'm not busy enough with that, I'm also currently completing my doctorate in AI, with a focus in Gen AI, through Golden Gate University.
Akanksha Bilani:
That's incredible. Kevin. And exactly to the theme of what we want to talk about today. What are some of the most exciting projects you've been working on lately?
Kevin Derman:
What comes to mind is United Airlines. Which is a great client of ours, and they wanted to really try and innovate on Gen AI use cases, and I love the way that they did this. They had an innovation day internally and they divided up all of their teams into 30 different units that were going to look at different projects. They developed 30 Gen AI use cases, and each team got a chance to present these use cases, and they eventually selected two that felt they were meaningful and they wanted to expand on further. The one was called EFHAS, E-F-H-A-S, which stands for Every Flight Has A Story. What EFHAS does is it provides a story to their clients, us, the passengers and it tells them why that flight was delayed. And they were able to take it from 15% of their flight delays, were able to initially get a story out to the people, to over 50% with this Gen AI solution.
They still put a human in the loop, which is very important, and I really liked that they did that. But it's been incredibly successful and I've experienced it myself. I was waiting for a flight and sure enough, that flight was delayed, and I got the message. And the message explained way more than just telling you the flight was delayed.
And it gave me a feeling of, ‘Oh okay, I understand.’ So here's AI doing something that makes it feel that that company is more human, which is great. The other, product that they did was also honing in on the humanity side of it by ensuring that their operators and their call center who are answering emails from irate clients, somebody who's upset that they've lost their luggage, etc., all of these different scenarios, and it gives it recommendations for being far more empathetic to the client and acts as a copilot for them in responding to those emails.
So both of them were very successful for the client. They were delighted. And it's a great story because it's from ideation right through to production. And we co-created with them. So we were involved in those teams in doing this. And of course, on the back end, if you go to the actual infrastructure that's happening there, it's built on AWS and they are sending those requests through to CPUs for those inferences as well.
Akanksha Bilani:
That's amazing. I know Intel has been working very closely with you, as well as your team at Slalom, Kevin, to really talk about some myths that you guys are busting right now being leaders in that, as customers or partners are thinking about building with Gen AI, how do we make sure that that service through data, as well as Gen AI platforms, actually dissipates with cost effectiveness and performance that can help with happy customers for United.
How about you tell us a little bit about the CPU on AWS story a little bit.
Kevin Derman:
The market is moving at such a rapid pace, and the innovation is happening at such a rapid pace that CTOs, people who are involved in choosing the cloud infrastructure, they need to keep up to date with what's happening. And CPUs, in particular, Intel, are doing wonderful things. So if you think about your Xeon 6 processors, and you just look at the difference in an inference cost.
So a typical use case of somebody going in and typing a prompt into a Gen AI chat bot and getting a response, there is no need for that to actually go through a GPU. That can be adequately served through a CPU, 33% faster. And that's research that has been done. And like 280% cost reduction compared to a GPU.
Every client is trying to manage costs, especially with AI. And especially in an environment where they don't know what is the extent that the solution that they put in place is going to be used? Is it going to be a thousand people that are going to use it, or is it going to be, we turn this on and in a week's time we've got 2 million users on the system and our costs have suddenly skyrocketed.
So there are a couple ways of dealing with that, using Amazon Bedrock, and being able to, use it in a serverless environment, can scale. But if you know that you're utilizing a certain capacity, rather go and spin up an EC2 instance where you can actually choose that Xeon 6 processor. And, you know, I am getting the best costs that I can for those inferences at that particular time.
Akanksha Bilani:
I love that, and thank you for thinking outside the box, Kevin, on ensuring that we are customer-focused first and we know what the power of silicon can drive. And absolutely, you are right. Inference is inference in training. If you look at both of those packaged environments in an AI platform, if you're not making sure that you optimize for inference towards production, even training is going to be a lot heavier and clunkier for customers to deal with that.
You guys at Slalom have cracked the code on what's the right thing based on data as well as ingestion with the power of silicon. So I think you've connected all of the dots in the right way. On the same path of, you know, thinking vision, and thinking what's nice and right for the industry in the next, say, 5 to 10 years...
Your vision as Doctor Kevin, what do you see is the next trend of innovation or technology that's going to disrupt our industry today?
Kevin Derman:
5 to 10 years, I see the biggest impact that we're going to have is that the technology that we're dealing with today, the AI technology, is going to start moving into the physical realm. In our lab, for example, I have a humanoid robot, in that lab. We are busy training the humanoid robots to be able to put it into a warehouse where it can walk around at night looking for health and safety issues and reporting back on those.
So the traditional LLM models are moving into the real world models, where robots can then understand the real world that we're in. This humanoid robot has 3D lidar. It has depth perception. So it can completely interact. It has an LLM that it backs into, so you can have a discussion with it so you can ask it to do things.
We're going to see far more robotics coming into the market. And it's going to become commonplace. I'll give you a time frame. I think that's within the five year timeframe.
Akanksha Bilani:
Amazing.
Kevin Derman:
The second one is human augmentation. And this is an area that I'm really passionate about. And I am fascinated by it. The brain computer interface world. If you think about companies like Blackrock, like Neuralink, like Synchron, and the work they are doing. They are being able now to train up LLM models to understand the brainwave patterns so that somebody who is paralyzed can think about raising their arm and a robotic arm will raise and can pick up a drink for them.
If they can't speak, they can think about what they want to say, and a computer will be able to say it. And that is advancing at such a rapid rate. And it's absolutely fantastic. If you're looking in that ten year period, I don't think it's only going to be the medically or physically disadvantaged people that will use these brain computer interfaces.
It's going to be offered to you and I, and we will be able to have this implant where we will have access to these LLM models. Instead of being augmented at the end of our arm with a phone, the augmentation will just be in our head. So I think it's incredibly exciting. And that's the world that we are heading into in a very, very fast timeline.
Akanksha Bilani:
That sounds so exciting. The vision's amazing, Kevin, I look forward to the future with you. Well, that wraps our show today. You've been listening to Rewired, the podcast celebrating innovation on the bleeding edge, brought to you by Intel, AWS, and Slalom.
Check out our show notes on the Rewired page at www.rewired-podcast.com, or wherever you listen to podcasts, for links to learn more about what we discussed today.
I'm your host, Akanksha Bilani, I would love to extend a massive thank you, Kevin, for being here and sharing these awesome stories and vision with us.
Kevin, thank you so much for coming.
Kevin Derman:
Thank you, Akanksha! I’ve really enjoyed chatting with you.
