We recently sat down for a conversation with Kesha Williams to discuss her journey from "playing on the computer" to machine learning and recent events and activities she's supported—like Kesha leading a talk and a three-part "Machine learning (ML) for Social Good" training series with Women Who Code, that culminated in a hackathon which challenged participants to apply their ML skills to help address challenges caused by COVID-19.
A woman with dark skin and black hair looks at the camera with a slight smile on her face
What was your journey into tech?
I actually started coding in high school, and I didn't even realize at the time that what I was doing was called coding. I just called it "playing on the computer." As it turns out, I was actually writing computer programs using the BASIC programming language. When I was in the 11th grade, I went to a six-week summer science enrichment training program at a local college where I learned more about computers and chemistry. That experience really sealed the deal for me, and I knew that I was going to have a lifelong love with computers and technology. When it was time for me to head off to college, it was an easy choice to choose computer science and math for majors.
I went to Spelman College in Atlanta, and I never left. My first job in tech was with the National Security Agency. That introduced me to web development. And from there, I can just say the rest is history. I've built a 25-year career on doing Java web development, serving as a software engineer, a technical lead, a software engineering manager. Then three or four years ago I started dabbling with machine learning and found my second love. It's been a journey, just enjoying learning new things.
What first got you interested in machine learning? What piqued your interest?
When I first heard about machine learning, I found the technology to be exciting and scary at the same time. I had this preconceived notion that I needed to be a research scientist working on my PhD in a lab somewhere to even really understand machine learning or apply it in the real world. Because of that, I saw it as a challenge. The way I'm wired, I love exploring and playing on the computer and learning new things. What I try to do is demystify the technology and show people that it's not as complex and hard as people make it out to be. If it's something that I can grasp, I'll show you all of my lessons learned to jump start you to pick up where I left off and continue on.
What inspired you to lead this Machine Learning for Good hackathon with Women Who Code?
I'm always challenging myself to learn new things and try new things. With machine learning, there are so many different learning algorithms and so many different machine learning techniques. I feel like I'll never master everything, but I can continue to advance what I do know. So over the summer, I set a personal challenge for myself, and that was to use machine learning to battle COVID-19. And I thought, "It would be awesome if I could take others on this journey with me." And so I reached out to the founder of Women Who Code to talk through this idea that I had about having a hackathon and some workshops.
I kicked off the hackathon at the Women Who Code Digital Connect Conference and did a 45-minute talk about using machine learning for social good and that led into three workshops. I taught members about machine learning fundamentals, different techniques, how to prepare data, and how to train a model. By the end of the workshop, participants had the tools to create a model to battle COVID-19. And so at the end of the third workshop, the challenge to the attendees was to form teams, come up with a way that they could use machine learning to battle COVID-19, and then submit a project to a hackathon for some awesome prizes.
What was the response to the training sessions?
I wasn't sure how it was going to be received. I wasn't sure if it was going to be popular. I was just totally blown away by the number of people attending, the level of engagement, a lot of questions, so many questions that I couldn't even answer them all online live during the workshop. We had about 686 people register for the three workshop series, and we had 211 people register for the hackathon with 52 teams. So there definitely seems to be a huge appetite for just machine learning in general.

We had people from literally around the world—Australia, the UK, India, Nigeria, Sweden, Italy, Ireland, Mexico, Singapore, Brazil, and the U.S. People from so many different time zones participating in the hackathon, the workshops and asking questions. I was surprised at the reach of Women Who Code. We also had a wide range of technical expertise, from people new to machine learning and coding to those who are experienced machine learning engineers participating in the workshops and the hackathon.
The hackathon results were announced on August 19, can you tell us about the submissions you saw and the winning team?
When watching the presentations, I was very impressed with and so proud of the teams. It was evident that they had taken everything I had taught them and used it as a basis to springboard their learning and incorporate new things that I hadn’t taught them. As a mentor and instructor, it is always my goal to set a strong foundation so that people can quickly grasp and easily master more advanced topics. I saw submissions that used machine learning to develop vaccines against COVID-19, predict risks of infection, classify pneumonia as COVID-19 or bacterial, categorize the risk of catching COVID-19 in a certain area, and determine why Black communities have a higher risk of death. The winning team used machine learning to predict whether a patient would need intubation with the goal of managing intubation supplies and reducing shortages.
I’m most excited about the fact that all teams will have an opportunity to continue developing their applications. While the winning team received $25,000 in AWS credits to continue with development and a 1-year free membership from A Cloud Guru to continue their learning, the other six teams received $5,000 each to continue development. I've spoken with each team since the hackathon concluded, and they are continuing to develop their applications.
Why do you think machine learning can help fight this pandemic?
Machine learning gives you new insights in your data. It helps you uncover relationships that you might not have noticed before or you might not have considered before. For example, I challenged myself to create a model that uses kidney function to determine or predict COVID-19 outcomes. Early on, we saw that kidney disease was associated with death in COVID-19 patients. I created a model that looks at kidney function to predict outcomes and find new relationships between that and COVID-19 that might not have considered before.
With machine learning, there are so many different use cases that we can solve for to help battle COVID. You could use computer vision, which is a form of machine learning, to determine if people are social distancing, to determine if someone is wearing a mask or not, to predict the spread. That's one of the projects that we saw in the submission. You can use it to predict outcomes and hospital capacity. There are just so many different ways that we can use machine learning to help battle the pandemic.
You’re part of the AWS Machine Learning Hero program that recognizes a select few individuals who go above and beyond to share AWS knowledge. Can you tell us a bit why you're partnering with AWS?
When I started my machine learning journey, I started with AWS. AWS really helped me go from a Java developer to a machine learning practitioner in no time. Just the AI and the ML services like Amazon SageMaker, it makes very complex topics approachable. When I was learning, the services that I chose met me where I was in my journey. When I was just first starting out, I wasn't ready to really understand the deep technical details of how machine learning worked. And so certain services would abstract that way. And then as I began to learn more and more, and I wanted to go deeper and deeper, there were services that allowed me to do that. Because of my journey, I wanted to introduce others to AWS. AWS gave me all of the tools and all of the support I needed to just bring some of my innovative machine learning ideas to life. I wanted other people to experience that. It was really a no brainer for me to partner with AWS.
So what's next? You're a wonderful evangelist for machine learning. How are you planning to continue to help drive and raise awareness?
I will just continue to do what I've been doing. I enjoy speaking at conferences and sharing my lessons learned and teaching workshops, authoring content on A Cloud Guru, and being a champion for AWS machine learning, Amazon SageMaker, and the community. I'll definitely continue to do that. I want to continue to introduce new people to these technologies.
For upcoming events and trainings on data science, visit Women Who Code.