In this interview, we speak with Salil, one of the Machine Learning Engineers at Armorblox. Learn more about what it's like to be a Machine Learning Engineer in cybersecurity.
Welcome to Blox Life – an interview series that shines a spotlight on our team members by chronicling their daily work, highlighting their achievements, and sharing their insights with people who wish to pursue similar careers.
Meet Salil Kanetkar, one of the Machine Learning Engineers on the Armorblox Engineering team. Salil has been with us for almost 3 years now, playing an integral role in building the technology and deep learning models that have propelled Armorblox as a top email security solution.
Q. What made you want to pursue a career in cybersecurity?
I didn’t set out to pursue cybersecurity in particular, but I did want to pursue machine learning. I gravitated toward cybersecurity because there is so much opportunity in this space that is under-explored.
There is also the added bonus that I get to use the newest, coolest machine learning algorithms while at the same time knowing I’m working to help protect people from cyber attacks.
Q. What is challenging about detecting fraud in email?
The most challenging aspect is its ever-evolving nature. Cybercriminals move fast, and we have to move faster. You have to be in sync with the newest trends or you will fall prey pretty quickly. With new scams, our models aren’t yet trained to detect them right away, so we have to adapt very quickly by changing our models to catch them to keep our customers protected.
We saw this with COVID-19 where there were new scams that were popping up overnight. This was also true for the Ukraine-Russian war, where we saw a huge increase in Russian fraud bots. Because of sanctions, Russia took a very aggressive stance in cyberspace and deployed attacks to retaliate against the US.
Q. What about your role excites you the most?
For me, the coolest part is having access to the most cutting-edge technologies of the day. Every month there is a new state-of-the-art algorithm that I get to learn about and work with. For example, there is a model called DALL-E that is next-generation AI. This technology is almost on the same level as a human in terms of how smart it is. It is trained to create images from text captions and it is very accurate. The fact that we get to use technologies like this is so cool.
Every month there is a new state-of-the-art algorithm that I get to learn about and work with.
Q. What are some features you worked on and launched recently?
Recently, I worked on a feature that allows Armorblox to detect fraudulent websites and landing pages. We’ve expanded our capabilities to go beyond just looking at the metadata and text on the landing pages where victims are being directed during phishing scams. Now we can look at a screenshot of the web page to detect fraud, which gives us another dimension. Adding this feature helps us catch a much higher percentage of fraud with greater accuracy.
I’ve also been able to work on our technology for detecting fraudulent payment requests. Catching these types of scams is so important because there is a very real risk that the person on the receiving end could lose money.
Q. What does a 'day in the life' look like for Salil?
Our team is big on supporting one another. We start each day with a stand-up meeting to discuss progress, answer questions, and ask for help if needed. Then we map out our daily tasks for that day. The meetings are very collaborative. We take this time to really help each other solve problems.
Then the rest of the day we get to work on building out new features or improving existing features. We do regular team meetings where we discuss products and provide a ton of feedback so we can make improvements. We also have a weekly book club to share the latest blogs and articles on computer science and AI, and we meet with the Data Analyst team to learn more about any new scam types they are seeing so we can strategize on how to prevent them.
The meetings are very collaborative. We take this time to really help each other solve problems.
On Fridays, we have a Detection team happy hour where we can hang out and catch up on life – the important things that happen outside of work. Sometimes we also do team meet-ups in the Bay area.
Q. What has been the most rewarding aspect of your role?
It is very rewarding to see my work moving directly into the live product. It is a great feeling to see end-users using your product, and not only using it but also providing positive feedback about it.
Sometimes at large companies, you never get to see this or it takes a very long time. But because Armorblox is a start-up you have a direct line of sight into your impact. Within a few months, you can see a feature you created go live. Getting that feedback from customers about how the product has made their life easier is rewarding.
It is a great feeling to see end-users using your product, and not only using it but also providing positive feedback about it.
Q. What has been your greatest learning lesson?
My greatest lesson has been that working together is so important and vital to creating strong, smart products. Great products are not the result of one person.
When you're on a team, it's very important that everyone moves forward, together. You need to always ensure that everyone has the same goal and everyone has the tools they need to move forward with that goal. If in a team, someone is not being helped, then the final outcome won't be as good as it could be. If there is a lack of communication and/or coordination, that will work against you. At Armorblox, our team has done a really good job of working together to meet a common goal.
Great products are not the result of one person. When you're on a team, it's very important that everyone moves forward, together.
