Employee Spotlight: One Minute with Maria Pospelova

Maria discusses the journey that led her to data analytics for cybersecurity and offers advice to aspiring female data scientists.

Mario Pospelova HeadshotInterset’s data scientists are constantly pushing past boundaries to deliver new, innovative ways to detect threats. With deep expertise in big data, analytics, and data science, Data Scientist Maria Pospelova takes a dominant role in the development of Interset’s technology, which helps to detect security threats before data is stolen. Maria is an exemplary innovator, recently having developed a new automatic NetFlow classification method that can autonomously determine traffic is HTTP(S), SSH, or DNS, without explicit information from interim hardware routers.

Prior to Interset, Maria worked on machine learning solutions for big data problems at Carleton University’s Parallel Computing Research Lab. Prior to Carleton University, she worked at Bedarra Research Labs on an Interactive Collaborative Analytics Environment product, where she had an opportunity to experience both front-end and back-end development. Maria is also a published researcher as co-author of “Automatic, On-Line Tuning of YARN Container Memory and CPU Parameters.”

Maria holds a Master of Computer Science with a specialization in Data Science and a Bachelor of Computer Science Honours with Highest Distinction from Carleton University. She is the first alumni of Carleton’s Data Science Institute, completing a two-year program in half the time with a Google scholarship. A continuous learner, Maria recently achieved CompTIA Security+ certification.

Q: How did you come into this field of work?

A: My path towards data science wasn’t straight or conventional. In the field of data science, my love for perplexing brain-teasers meets my pragmatic nature and desire to save the world—one puzzle at the time. Today, AI and machine learning (ML) are seen by many as the panacea that will cure all of the world’s problems. Unfortunately, these data science techniques have their limitations, and it’s important to understand when, where, and how to leverage them for the best impact. In cybersecurity, AI and ML are highly effective and, in my opinion, the only feasible solution to the growing avalanche of security threats.

Q: What is the most inspiring part of your job?

A: Catching the bad guys. At the end of the day, our algorithms, mighty big-data-handling architecture, and clever user interface come together for one single purpose: to stop malicious actors. That “gotcha!” moment—the one that inspires so many investigators and bounty hunters—fuels my motivation. I am particularly fortunate that, in addition to the wins I see first-hand during new model development and POC investigations, I also get to hear stories down the line from our customers about catching cybercriminals in the act with Interset’s platform. This is why I do what I do.

Q: What is it like building models for cybersecurity data analysis compared to building models for other types of data sets?

A: Cybersecurity is a constant arms race. Malicious actors have access to the same technology, knowledge, and resources. Fighting against AI and ML algorithms that are used for nefarious purposes without the newest data science and engineering technologies is like bringing a knife to a gunfight. And just a few models will not do; you need the latest and greatest models that produce fast, reliable, and interpretable outcomes that can immediately be turned into action by security teams. All of this is exacerbated by heterogeneity, variety, and volume of the data, as well as the constant metamorphosis of targets (no pressure at all!). This is what makes cybersecurity the most challenging—as well as the most suitable—application for data science.

Q: Any advice for aspiring women data scientists?

A: Go for it. Be curious and be brave. The confidence gap is at the very foundation of the gender gap. And the direct correlation between confidence and success is indisputable.

As is the case in many industries, diversity in data science is a big problem. If someone doesn’t look like a data scientist, she (or he) is unlikely to be seen and respected as one—by others or by herself/himself. This perception issue so easily undermines confidence, creating a vicious cycle. It takes courage to overcome the “forcefield” of assumptions. When I took a job as a teaching assistant (TA) at university, I was amazed at how often the appearance of the students didn’t correspond with the quality of the work they produced. The not-your-typical programmer frequently produced the most outstanding work. And these “unconventional” but brilliant minds filled me with inspiration and motivation to persist as a woman in a male-dominated field.

It’s all too easy for women to underestimate their own abilities and turn away from male-dominated industries, and we see it happen so often. Recent research suggests that after our 40th birthdays, the confidence gap between women and men disappears—but who has time to wait? Be confident now—it’s your well-deserved right. Actively work on improving your confidence just as much as you work on expanding your data science skills. It will make all the difference.

Rapid-fire Round

Q: Coffee or tea?

A: Both

Q: Classical or hip-hop?

A: Rock

Q: Tropical beach or urban cityscape?

A: A beach near a vivid city would be perfect—like Barcelona! If forced to pick one, I’d go for a cityscape.

Q: Electric toothbrush or regular toothbrush?

A: Electric

Q: Ice hockey or curling?

A: Figure skating

Be sure to connect with Maria on LinkedIn.