From Barrier Breaker to AI Innovator
Autonomy and AI Research Engineer Maria Angelica was just a young girl on her first flight to the Colombian Math Olympics finals when her dream to work in the aerospace industry was born.
Two decades later, we spoke with the ambitious research engineer to learn how she overcame obstacles and joined the Artificial Intelligence team at Lockheed Martin Rotary and Mission Systems.
Building a Strong Educational Foundation
Maria Angelica grew up in Colombia, where she studied electrical engineering at the University of Los Andes in Bogota, Colombia. Right after finishing her undergraduate degree, pursued a master’s degree in electrical engineering and computer science with a focus in controls. Maria Angelica taught herself how to do her master’s thesis which was an application with reinforcement learning, a technique that sparked her passion for AI.
Despite having two degrees, it was still a difficult process to kick start her aerospace career in the United States. Her early career consisted of joining a work travel program at an amusement park in Pennsylvania, before getting a summer internship at Collins Aerospace. After graduate school, she made the bold move across the world to work for a manufacturing company in Japan, but later returned to the U.S. where she landed a job as a software engineer on the Sikorsky flight control team in Connecticut.
With the help of Lockheed Martin’s tuition assistance program, she received a graduate certificate in robotics engineering with a specialization in autonomous vehicles form Worcester Polytechnic Institute.
While working at Lockheed Martin, Maria Angelica also achieved another one of her dreams. She attended MIT and recently earned a professional certificate in Artificial Intelligence and Machine Learning.
Reaching the Finish Line
Maria Angelica’s journey to her dream job wasn’t without challenges. She ran into several citizenship restrictions on the roles she pursued, but she never gave up.
She knew all she needed was one opportunity. That’s when she found Power Up, a Lockheed Martin program that gives employees the chance to enroll in stretch assignments and demonstrate their skills. After receiving mentorship from leaders across the business, she landed an AI stretch assignment role and six months later the team opened up a full-time position.
Today, Maria Angelica is programming and learning every day in her dream role. She enjoys being able to work remotely while being a very present mom to her young kids.
Outside of her current role, Maria Angelica is very involved in the Hispanic Organization for Leadership and Awareness (HOLA) business resource group, where she gets to connect with employees from different leadership and functional backgrounds. As the Connecticut co-lead she was able to organize visits to local elementary schools to introduce young children to the world of helicopters.
Keep reading to learn more about Maria Angelica’s role as an AI Research Engineer.
Q&A with Maria Angelica
Can you describe your role as an A/AI Research Engineer?
I work for Lockheed Martin’s Rotary and Mission Systems AI team. I focus on the research and development of technologies that enable and advance semi and fully autonomous systems for both defense and commercial products. Currently, I work in an Internal Rapid Application Development (IRAD) where we are focused on missile defense. We create and benchmark algorithms to solve the sensor and weapon target assignment problem with moving targets using machine learning. I am interested in the decision-making algorithms and often find myself playing with reinforcement learning, genetic algorithms, and attention-based networks (LSTMs and transformers).
What is a typical day like for you?
My team uses agile, every spring we plan our stories and I know my goal. Because I work as a researcher, some of the tasks are open ended since we are creating new products to expand the market and develop new solutions using AI. It’s amazing to have the opportunity to push the envelope. Here you are creating something totally new for a company that solves real-life problems. That’s what I love about my role, I never have repeatable tasks. I never have to do the same things over and over. Some of my daily tasks are reading and writing code, reading papers, and attending meetings. My favorite part of the day is playing Jeopardy or LinkIt with my team at the end of the scrum meeting.
How can someone pursue a career in AI at Lockheed Martin?
At the Lockheed Martin Artificial Intelligence Center (LAIC), there is a program where if you don’t know programming, they will teach you how to program in Python, which is what most AI applications are developed on.
They have phase one and phase two. In phase one, you get access to courses for the basics and there is supervised and unsupervised learning. Then there is phase two where you actually get to solve a problem in your current role using AI, and you get access to mentors from all over the company to help solve that problem. That’s a good way to network because you never know which mentor can recommend you for a role. It also showcases a real-world application that you were able to solve by using AI.
If you are not a Lockheed Martin Employee, I recommend having a strong experience and academic background in AI, build up your portfolio, be on the lookout for positions, and try to attend to conferences that Lockheed Martin goes to such as the Society of Hispanic Professional Engineers, Society of Women Engineers and National Society of Black Engineers.
Why do you think AI is so valuable in the aerospace and defense industry today?
We are in the digital era. We are talking about intelligence agents now that are making decisions. So, before we had a human telling the machine what to do and where to go. But now, we need developed planning and decision-making abilities based on the data in our surroundings.
AI is so powerful because systems are huge nowadays. Now we have platforms with incredible amounts of sensors and weapons. What AI can do is, in real time, process data to make good decisions in seconds, once a good model is deployed.