Exploring AI with Kirkwall

Written by: Brady Drake

What is Kirkwall?

Kirkwall, founded in 2022, is a company that has evolved from focusing narrowly on cybersecurity for drones and autonomous systems to providing a more holistic monitoring solution for robotic and industrial systems. This transition was inspired by feedback received during customer interviews conducted as part of the NSF I-Corps Program.

Kirkwall’s solution acts as a “check engine light” for robotic and industrial systems, offering a broad spectrum of monitoring capabilities that include both cybersecurity and physical maintenance and their Machine Learning algorithm allows for automatic shutdown due to security or safety concerns.

What problem Kirkwall is trying to address?

Kirkwall addresses the need for comprehensive monitoring of robotic and industrial systems, beyond just cybersecurity. They recognized that companies were looking for solutions that could alert them to a variety of potential issues, including the risk of machine failure due to physical defects or maintenance needs, in addition to cybersecurity threats. This broader approach aims to ensure that systems operate smoothly and safely, minimizing downtime and protecting against both digital and physical threats.

About Devii

Devii, based in Fargo, is an instant API engine and online database hosting platform. It’s designed to help software engineers develop better code more efficiently by automating some of the most time-consuming tasks that backend engineers face during digital transformation projects. By automating the repetitive aspects of software development, devii enables engineers to focus on writing more meaningful code.

How does Kirkwall address the problem?

The company utilizes advanced technology, including machine learning and an automatic API engine powered by devii software, to efficiently pull in and analyze data from various systems. This technology enables quick integration of new data feeds and the application of machine learning for anomaly detection. By identifying what normal operation looks like for a system, Kirkwall can flag abnormal behavior.

Where is the Company heading?

Already up to three full-time employees, things are starting to pick up steam for Kirkwall, which acquired Heath Company in August of 2023.

“We have a lot of good stuff in the pipeline,” said Co-Founder William Cromarty. “We haven’t officially opened a venture capital round yet, but we’re probably not too far off—and things are looking good for that. Recently, we did our software launch out in California, and that went very well. So now our software is commercially available, and we’re onboarding a handful of clients. What’s nice is that we have a couple of big names in there. We’re doing a rollout for a Buffalo Wild Wings location, which probably sounds surprising, right? Buffalo Wild Wings is a little outside of what you normally think of for autonomous systems, but we’re actually getting a lot of interest in the smart building automation sector. So, Buffalo Wild Wings is an early pilot focused on monitoring essentially everything in a building—everything from temperature sensors to electrical, all of that. Pavewise, a startup out of Bismarck, is another early client of ours. This year is going to be a big push for how many users we can onboard and how much monthly recurring revenue we can achieve. We actually might be buying an industrial space in Fargo relatively soon as well. I’ve also spending a lot of time in Sioux Falls lately as part of our expansion this year into Sioux Falls and Grand Forks.”

Who is William Cromarty?

William Cromarty is the CEO and founder of Kirkwall, president and CEO of Heath Company, executive director of the Museum Incubator, a Corporate Board of Advisors member for the Research Institute of Autonomous Systems, a Board of Directors member for The F5 Project, a board member for Cass County Government, and a former CIA officer, among other things.

Who is Ujjwal Adjikari?

Ujjwal Adhikari is the co-founder and CTO of Kirkwall. He is a graduate of Emerging Digital Academy and he “draws on a diverse range of full-stack coding expertise with specialization in CRM software development and application of Machine Learning to autonomous system diagnostics.”

Kirkwall is Hiring!

Kirkwall is currently hiring for five paid internship roles located in North Dakota. As an added bonus, all interns will have the opportunity to use Kirkwall’s skydiving stipend and will receive basic training on the company flamethrower.

“Incorporating both the flamethrower and the skydiving stipend into our company culture is Kirkwall’s way of selecting people with above-average risk tolerance and drive for adrenaline,” Cromarty said. “Entrepreneurship is frequently a highadrenaline activity with lots of unknowns, so we like selecting candidates who comfortably operate in that space as a way to reinforce that the pit in your stomach is often a sign of acceleration and progress rather than something to be feared.

How is Kirkwall exploring AI?

Proposal writing and Grant writing

“I’ve personally found it really helpful for proposal writing and grant writing. You obviously still have to go back and fine-tune everything, but I have found it really helpful for knocking out the first 80%,” Cromarty said.

Deciphering code and finding inspiration

“My Co-Founder Ujjwal is utilizing ChatGPT in a pretty interesting way,” Cromarty said. “One of the ways I’ve seen him be really effective with that, from a coding perspective, is when he finds an example online of code that could be the inspiration for something that we want to do. Ujjwal has found it very helpful to just paste that code into ChatGPT and have it explain what individual lines of code do. He is using it as a really great educational tool for new capabilities.

Edge Processing

“One other use case that has been interesting for AI and machine learning is edge processing,” Cromarty said. “This is one that I’ve used in the past. And it’s one that we’re exploring, basically, for any kind of aerospace system. One of the biggest challenges is always bandwidth. Traditionally, what people have done in the past is pull down all the data from a sattelite and then analyze it on the ground. One of the really interesting things that AI and machine learning are opening up is allowing the system to decide whether the data it’s collecting is useful or not. What that means is that you can send down only data that you consider useful. The satellite can determine the info on its own that it thinks is going to be relevant to you, which means that you’re not wasting a ton of bandwidth, just sending everything down. All of a sudden, you can focus on what’s important. Everything being sent down is of high value. This still takes testing before you can trust it fully and there are times when you’re going to have to pull everything down and grade it as you go and give it a score on how good of a job it did so you can keep teaching it.”

Interface Development

“Ujjwal has actually been showing me some really interesting spin-offs of ChatGPT-4,” Cromarty said. “In the most recent launch, you can make custom models based on the training data you give to it. He’s found some really interesting spin-offs there that allow you to generate web pages and actual code based on what you ask it to do. He has built entirely new interfaces using this as experiments over the course of an hour. I’ve been pretty impressed with that, and we’re already putting that capability to good use.”


Learn More


Share This Article
Brady is the Editorial Director at Spotlight Media in Fargo, ND.