W&M’s Ted Hallum Connects Veterans with Opportunities

Meet Ted Hallum, MSBA ’20. This U.S. Army veteran and Senior Defense Machine Learning Engineer for Octo Consulting Group hosts The Data Canteen—a podcast that grew out of the Veterans in Data Science & Machine Learning (VDSML) community, which he founded.

He recently spoke with us about his inspiration to connect veterans interested in data science and machine learning, the remarkable speed at which his programs continue to grow, and the particular benefits and advantages provided by William & Mary’s Online MSBA program.

Excerpts of our conversation follow here.

“When I started my professional career as a soldier, I was an intelligence analyst. I did five years in the Army. When I separated from service, I had the security clearance and the skills of an intelligence analyst. I went to work as an intelligence analyst in defense contracting, primarily still supporting the Army, just in civilian clothes instead of a uniform.

Fast forward eight years to when I first encountered data science. I was working in an intelligence agency, doing intelligence analysis around other intelligence analysts.

At that time, nobody in my physical sphere at work, nor in my LinkedIn community—no one that I knew—had touch-points with data science. Nevertheless, I started to get interested in all its implications for intelligence analysis, and for the world, more broadly. I started my data journey alone—from being interested in it to reading about it, getting excited enough to apply to grad schools and, finally, starting a graduate program. At many points, it felt like being in a pitch-black room, fumbling around, looking for light switches.

I had this idea that started on the back burner, probably midway through the Online MSBA program, that there had to be other veterans interested in these subjects out there. It was just a matter of finding them and bringing us together.

Then in the future, when other service members decide to separate from the military and maybe pursue a path in analytics, data science or machine learning—the whole so-called datasphere—those of us who are a little bit further along, who have done a graduate school program or a boot camp, or have three or four years in the industry, we can help one another. I felt that it could be a much more linear path for those coming in my wake. So that was the catalyst. That planted the seed of desire to start a community like this.

I knew it would be niche, but I thought, “Maybe if I kicked off a LinkedIn group, I’d eventually find the other 40 or 50 veterans who are also interested in this kind of thing,” and here we are. We're just north of 800 people on LinkedIn groups. 800 is not three million, so we're always going to be a small group, but there's much more interest than I initially anticipated.

As it turns out, many new folks who are just coming out of uniform are asking, ‘What graduate school program should I pursue? What's the difference between data analytics and business analytics and data science? Where does machine learning fit into all that?’ There are so many question marks. It’s all opaque to people who aren’t in the datasphere.”

“I graduated from William & Mary's Online MSBA program in August of 2020. I just thought, ‘I’ll create this LinkedIn group and veterans who are interested in data will join it, and then we can communicate and help one another.’ I was initially naive about what successfully launching the community would entail. If you remember the movie Field of Dreams—‘If you build it, they will come.’—that was my initial train of thought. Today, the LinkedIn group is still a big part of our community, but I quickly realized that VDSML would require more cultivation, leadership and infrastructure than just a LinkedIn group. And so, since it was my idea, I actively took the leadership role.

Only a few weeks later, someone said, ‘We should have a podcast because that's a great way to disseminate information. We want to highlight the strides and striving of veterans in the data space. A podcast is a great platform to bring in veterans who are succeeding out there.’ That does a couple of things: It shows the world the impact veterans are making in the data space, and it holds them up as a template.

If you're a new veteran coming out of the service, and you need to know how to go from where you are to being a principal data scientist or the director of a data science program, the answer is you listen to VDSML’s podcast, The Data Canteen. You see a fellow veteran who used to be a sergeant coming out of the Army, who has gone through the steps of an upskilling and worked up the career ladder, and then you do what that person has done. You don't have to reinvent the wheel. You can follow a template. Providing those templates is one of The Data Canteen’s primary objectives.

We also launched a mentoring program, which was really a natural evolution. It quickly became apparent that there was a need for it. Then, as more and more people joined the community, especially people with four and five years of experience, it also became apparent that we could offer meaningful mentorship to new people.

The great thing about veterans is that they naturally have a heart of service, so it wasn't as hard as you might think to generate interest and get people to help more junior folks along.”

“There's the mentoring program and the podcast. For our premium-level members, we offer a Slack workspace to facilitate the exchange of information in topic-focused channels, and a Discord, which is our virtual hangout space. Once a month, we do a virtual happy hour there, where people can come together and meet.

In Slack, we have an app called Donut, which automatically matches people for virtual coffee once a week. It provides a natural and routine means for members to get introduced, meet, and establish relationships with other people in the community.

One thing coming online soon is VDSML’s Projects Program. The idea behind it is that we have a lot of members who are early in their data journeys. They’ve done some upskilling—maybe a boot camp or a graduate school program—but hiring managers are looking to see real-world projects.

We're identifying veteran-owned small businesses and veteran service organizations that could significantly benefit from data science but do not have that organic proficiency, and probably couldn't afford or justify the expense of having a full-time data scientist on staff. Then we’ll put together small teams of our members who need to get real-world projects under their belts. A more senior member of the community will be assigned to each team and serve as its project manager. These teams will consult with their assigned veteran organizations, develop a deep understanding of the business problem to be solved, and respond with an appropriate data science solution—pro bono.

Everybody wins. The veterans’ organization will benefit from our membership’s data science expertise. Our members will get to work on real-world projects which they can spotlight in their GitHub portfolios. So when the hiring manager asks, ‘Have you ever worked on a real problem that made a difference for a real business or organization?’ they can say, ‘Absolutely, I have, and I did it as part of a distributed remote team, with version control, and here's a link to it on GitHub.’

VDSML has another program in the works that will probably come out at the end of this year or the beginning of next year. We’re working to establish talent pipelines with companies that have initiatives to hire more veterans. If a company wants to hire veterans who are data scientists, but they don't know where to find them, they can come to us. We'll have people who have already done boot camps or programs like William & Mary's Online MSBA. They'll have already participated in our Projects Program and they'll have a real-world project to show.

These veterans will be perfectly positioned to go into an interview and knock it out of the park. So there'll be a linear path from participating in our community to securing gainful employment in analytics, data science, machine learning engineering, or elsewhere in the datasphere.”

“To a certain extent, it's just statistical probability at play. I think people underestimate the size of the military. Every year, approximately 200,000 service members exit the military. That's just the people who leave.

‘Data science’ as a term was coined around 2011. The data science movement was very small then, but it picked up steam substantially by 2016-2017. Now, it's gotten to the point where, if you mention data science and machine learning, most people have some idea of what it means. They’ve at least heard the term. It's not completely foreign anymore, and I think that's just going to continue to be more and more of our world and economy.

Data science is going to fuel every form of business. So, think about 200,000 veterans exiting the service each year. Over the next five to 10 years, there's going to be an increasing slice of that 200,000 people who—because of their interests, their skills, their background, and changing needs in business and global economic realities—will funnel into data science and machine learning. And we'll be the community waiting to help them along.

Once you've done a tenure in the military, you inevitably have some unique skills in terms of determination and resilience. Service members begin cultivating these skills almost immediately upon entering the military. During basic training, every service member fails. This is by design. No one goes through basic training and executes every task with a perfect record. One of the main points is for you to fail, often to fail miserably, and not give up because you’re being prepared for the potential realities of war, where things will go wrong. No plan survives first contact with the enemy. So what do you do? When everything is in shambles and not going according to plan, what do you do? You have to figure it out, and I found that at play when I went through the Online MSBA program at William & Mary.

When I was working on an optimization problem at four in the morning and it was due the next day, and my Python code wasn't working and I had read the chapter of the textbook five times and watched all the lectures online 10 times and run my code 10,000 times, but it just wasn't working … what do you do? Well, you don't quit, you don't give up, and I didn't. I ended up graduating with a 4.0, but that wasn't because it was easy. Several times, I had an assignment due before the start of business on Monday. I worked on it all night, finally hit the Submit button, looked at the clock and it was nearly 5:30 a.m. I just brewed a cup of coffee and had to go to work. The United States Army uniquely prepared me to endure that.

I’ve learned that we're all capable of so much more than we think. I was forced beyond my imagined limits of physical strength and endurance in the military. For me, the Online MSBA program was the intellectual equivalent of that. It pushed me to technical and academic achievements far beyond those I would have ever thought myself capable of reaching.”

“From day one, looking at the school's website, I could tell the school was military-friendly. William & Mary’s Raymond A. Mason School of Business engineers programs to help veterans succeed.

When I was in the program, I was able to see firsthand that there's a decent percentage of students who were either active-duty military or veterans. That demonstrated that William & Mary’s passion for supporting veterans isn’t just marketing on the website. The university was actually enrolling and serving the academic interests of veterans.

This is a huge thing, and I’m certainly not alone in the veteran community in this regard: I had no STEM academic background.

I had done a couple of Coursera Data Camp courses—enough to solidify that data science is something I wanted to pursue. Still, I did not come from a STEM-focused undergraduate background, so many of the grad school programs that I looked at were immediately off the table. They had a long list of prerequisites and no mechanism in place to help you bridge that gap. It was essentially, ‘Find a community college somewhere, take this list of courses, then come back and we'll see what we can do.’”

By contrast, William & Mary’s Online MSBA program has four prerequisites. Not only is that doable, but they’re well-packaged. If you need those courses, they form your first semester. If you pass all of them with a satisfactory grade, you automatically matriculate into the core curriculum. It was very straightforward, and that was huge. I saw a lot of schools that didn't have anything like that in place. For the United States to remain globally competitive, we need to be putting people who want to be engineers into the express lane, not having our academic programs structured so that people are deterred from going into those fields. So William & Mary is definitely doing it right.”

“The whole program is completely applicable.

I use what I learned in the two machine learning courses and the artificial neural network course daily. I significantly benefit from the second half of the Online MSBA's artificial networks course—the portion that goes beyond tabular data with modules on natural language processing and computer vision problems. Where I work, the client base that we support primarily has computer vision problems. Getting that exposure to using artificial neural networks for image classification was huge. That was probably a prerequisite for getting the job that I have now. Moreover, computer vision is a perfect example of a skill that most MSBA programs don't cover, and a way in which William & Mary stands out at the absolute top of the pack.

Even as a machine learning engineer, I make regular use of what I learned from the Online MSBA program’s emphasis on communicating insights. My work focuses on taking machine learning models, helping to get them into production, and then monitoring and maintaining them over time. Data patterns in the real world drift, causing machine learning models in production to decay. When you're dealing with data science and machine learning, even as a machine learning engineer, you end up having to explain that stuff to people—to your immediate managers, to upper-level business stakeholders. We’re often meeting with less technical clients, but even then, everyone in the room must be able to understand why something is or is not working well.

If you know how to do everything, but you don't know how to get it out of your mind and make it understandable to an audience, chances are that your work, even if it's good work, is going to be deemed a failure. The communication piece was a benefit that I didn't anticipate. I’m really grateful I got that.

The William & Mary program has a heavy emphasis on statistics. Going through those courses gave me an invaluable statistical intuition. There are so many problems that I encounter where statistics are at play. From that intense immersion in statistics at William & Mary, when I encounter statistical problems, I have a sense, statistically, of what is going on and what's causing it. If I need to dig deeper into what's happening, I know how to do that, thanks to the coursework at William & Mary.

Often, the best way to communicate an idea to another person is with something visual. One of the William & Mary courses I took focused on data visualization and databases. I have found that I can waste an hour trying to write a thorough email or trying to have a conversation with someone, or I can just take the time to produce a good chart or graph and then I don't even have to say anything. Within seconds, they look at it and they know what I would have wasted a lot of time trying to communicate otherwise. So the data visualization part, combined with the general emphasis on communication, has proven to be huge.

Finally, while none of my jobs in the datasphere has focused on solving optimization problems, it’s been my experience that they arise occasionally. Moreover, when I’ve encountered them, they tend to be time-critical and vitally important to the business. There are a lot of data analytics and business analytics programs out there in which students don't get the exposure to solving optimization problems that William & Mary’s Online MSBA program provides. I’ve been thankful, multiple times over, to have had that training so that I could make use of it in those critical moments.”

Data science and machine learning are critical tools for organizations worldwide, and demand for a business analytics skill set will continue its global expansion in the coming years.

Maximize your data science credentials. The intensive 32-credit-hour curriculum of William & Mary’s Online MSBA program teaches you the necessary analytical skills and communication techniques to work with big data sets, machine learning and artificial intelligence—preparing you to succeed in this high-growth, high-impact field.

To explore your next steps on a career path in data science and machine learning, speak to a William & Mary Admissions Advisor today.