More Problems, More Solutions: Cyber Chief Scientist Explores Endless Possibilities

As chief scientist of cyber operations research, Les Servi says he “helps decision-makers frame and then make hard decisions.” What wasn’t a hard decision was becoming involved in MITRE’s longtime mentoring and work-study partnership program with Notre Dame Cristo Rey High School in Methuen, Mass. We recently spoke with Servi about the satisfaction of solving diverse problems, his projects at MITRE, and the fulfillment he finds in mentoring a new generation of leaders and innovators.

What makes MITRE great is the diversity of projects for a wide variety of missions. There’s dignity in all work—but for me it’s much more satisfying to make the world safer than it is to make toothpaste taste a little better or generate more web clicks. 

That is why I do what I do—plus some proselytizing to encourage others to apply optimization skills to national security problems. 

My current focus is cyber operations research and optimization. There are a lot of people who know a lot about cyber, and a lot of people who know a lot about operations research (OR), machine learning (ML), and optimization methods. The niche I’m trying to fill is where these fields intersect. 

The Poetry of Optimization

Some people think ML can do everything, but of course that’s not correct. AI is staggeringly powerful, but we shouldn’t pull out a hammer if we need a screwdriver. In fact, here’s part of a poem I wrote for fun one weekend on this subject (with the help of ChatGPT):

• Machine Learning, 
  though advanced in might, 
  Is not the answer to every fight. 
     
• Operations Research,
  it's not a foe, 
  But a valuable tool, 
  one should know. 
     
• So let us choose our methods with care,
  And ensure that the right tool is there
  For each challenge holds a unique demand
  And both machine learning and OR should be at hand.

In our sponsor work, we look at their systems to find the critical nodes. Our team’s solution algorithmically identifies a decision tree for how a system can best be attacked while accounting for some actions possibly failing. 

Thinking Outside—and Inside—the Box

My work in cyber doesn’t cover all I’m doing at MITRE, though. I’ve found that solving problems creates a cross-fertilization into other areas, because methods that crop up in one application can often be used somewhere else completely. 

I led a team that was recently issued a patent for reducing healthcare data collection burdens. It uses a new method that simultaneously accommodates imperfect initial data while exploiting hidden redundancies. The targeted domain of this patent is far from much of my work, but the underlying mathematics is not. Currently I’m working with a rock-star team on a fascinating problem involving algorithmically arranging boxes into a three-dimensional irregular space. Connections are everywhere.

Outside of MITRE, I’m very involved with the Military Operations Research Society (MORS). Last year as president, I decided the MORS theme for the year should be “analytics as a force multiplier”—that is, a tool not unlike other force multipliers.

Preparing Future Problem-Solvers

Growing up, I always liked puzzles, and as an undergraduate I was very interested in pure mathematics. Over the course of my education and career, I recognized the satisfaction of problem-solving in different application domains. I have more than a dozen telecommunication patents and even wrote a paper about pi.

This persistent curiosity comes up in our work with students at Notre Dame Cristo Rey High School (NDCR). For several years I’ve been mentoring a now-graduate of NDCR, Jhanny Jimenez Feliz. He’s so talented that from the start I just gave him harder and harder tasks to work on and never told him he was operating at an advanced undergraduate level. He always rose to the occasion. 

Jhanny possesses a very rare tenacity and resilience. In 2022, he was the first student in the school’s history to be accepted to MIT—and he’s interning at MITRE during his breaks. Right now, he’s trying to apply machine learning to some of our nation’s problems. 

Who knows what he’ll discover … and where those ideas will take him next?