This interview series shines a spotlight on Armorblox team members - chronicling their daily work, highlighting their achievements, and sharing their advice for people who wish to pursue similar careers. In this interview, we speak with Holly Pike from the Armorblox detection team.
Welcome to Blox Life! This interview series shines a spotlight on Armorblox team members - chronicling their daily work, highlighting their achievements, and sharing their advice for people who wish to pursue similar careers. We’re delighted to kick off this series by speaking with Holly Pike. Holly is a technical lead (NLP) on the detection team and plays a vital role in creating, fine-tuning, and deploying Armorblox ML models that protect our customers’ people and data. Take it away, Holly!
Q. Can you tell us something about yourself?
Hi, I'm Holly. I've been at Armorblox for a bit over 2 years now, and I came to Armorblox straight out of my undergrad. I actually started school as a psych major, so being an early engineer working on NLU at a cybersecurity startup was definitely not what I expected. But, it's also felt like a completely natural progression to get here and Armorblox has been a great place to land.
Q. What is your current role at Armorblox? What does a typical day (or a typical week, if you can’t fit everything in one day!) look like for you?
I'm an engineer on the detection team here. One of the best things about working on a team as broad as 'detection' is that you get to be responsible for so many different parts of the product. Every problem we're solving will have a different solution, from incredibly complex and finicky models to deceptively simple algorithms. In one week I might be doing a deep data dive to understand the new problem we want to solve, the next building and tuning a model, the next working on backend systems to get that model in front of our customers. No two days or weeks are ever the same, but one thing that does stay constant for me is plenty of reading to try to keep up with the amazing new research and projects coming out in NLU and ML. It's always a balancing act between research, data science, model building, and backend work, in the best way.
Q. Why did you decide to join Armorblox?
I feel like I have so many answers to this question looking back on the decision, knowing it was the best decision, but it came down to believing in the company and knowing that Armorblox was going to push me. In my first conversations with the team it was clear that everyone believed full-heartedly in the vision; I don't know who could see Armorblox through how they talked about it and not be sold. On top of that it was clear Armorblox was going to be reciprocally invested in me. I think the pitch when I was interviewing was that I'd learn as much in 1 year at Armorblox as I'd learn in 3 at a big company, and knowing how much I wanted to accomplish in my career it seemed like a perfect place to keep reaching for every opportunity. And, for the record, Armorblox has absolutely lived up to that interview promise.
Q. What is an accomplishment at Armorblox you’re proud of?
Wow, I could never pick just one. I'll cheat a bit and say I'm always proud that my work impacts our customers so directly, which I'm sure is a common feeling here.
Working on detection, there's an immediacy to what we get done and how the customer's next day will go as a result of it. Having our detection engine stop a sophisticated or interesting attack can make the day feel as great as missing the attack would have made it a security admin's worst day. Every time we push a new model that better protects our customers, it's an accomplishment in my mind.
Q. What’s one thing you wished you knew when you started your career in this role (at Armorblox or in general)?
I think it's important to remember that everyone is learning every day (and if you're at a company where that's not the case maybe don't be at that company anymore).
Sometimes it can feel like there's so much to learn you'll never catch up or keep up, especially at a startup where everything moves at warp speed. For me the best way to learn was to focus on how to think about the problem rather than focusing on the details. Figuring out how to figure things out is the best way to keep progressing in step with the progress of the field. In a nutshell, ask questions to help you understand the why more so than the what. Try to be comfortable not knowing and then learning. Be flexible.
Q. Do you have any advice for people who aim to pursue a career similar to yours?
So I joke pretty frequently about being a stubborn person, but I think it's been invaluable in getting me here now. None of the components of where I am are the easy option -- startups are hard, cybersecurity is a high stakes game, production-grade NLU is a very different world than research.
Thinking of these components separately I have different advice for each. Being at a startup was the best way to kickstart my career, I've never doubted that decision. On the cybersecurity side, don't be intimidated out of a field just because you're not an expert in it yet. A combination of an eagerness to learn and amazing teammates willing to teach can cover any perceived gap. It's also how you can diversify a field, by taking that step and becoming a bridge for whoever comes next. And finally, for ML or NLU just ignore any companies not willing to take a chance on new talent. With how fast ML is changing, having education or experience won't guarantee success, and being able to keep up with the new is only going to be more and more valuable.
As a whole I'd say to focus on the problems you want to solve. For me the combination of solving NLU problems that have a human behavior undercurrent, in a tight-knit, challenging, exciting environment has checked every box I was looking for in a career while being totally different than I had ever pictured.
We hope you enjoyed getting to know Holly a little bit better! For email security news and trends, unique email attack detections, and more Blox Life, subscribe to Armorblox email updates below.