Today’s blog features Jasmine Alkin, a Technical Lead on our Machine Learning Operations team and one of the many talented data engineers working at BenchSci. We appreciate her taking the time to answer a few questions about her experiences working on our machine learning models and the complex but rewarding challenges therein.

Why did you choose BenchSci?

I chose BenchSci because it feeds into two of my passions: one being genetics and biology in general, and the other being computer science. In pursuing my career as a data engineer, I’ve often worried that I’ll lose my connection to the biological sciences. Working at BenchSci allows me to not only increase my skills as an engineer, but also my understanding of how science and biology work—especially when I need to model those concepts within the work I do.

What’s been the biggest benefit of moving to the country since being remote?

I'm an introvert through and through. I find being around a lot of people can be very draining. I never realized how much of a mistake it was for me to live in a city until I moved out. My energy stores are so much higher up here—I feel better, I have more time, and I have much more energy for being active and engaging in my hobbies. I think that's the biggest difference. When it comes down to it, different people thrive in different environments, and living in the city just wasn’t working for me. Living in an environment I am better suited for has helped me so much.

What's the most interesting project you've gotten to work on so far?

It’s not necessarily what I’m working on now that I find most interesting—it’s the possibilities for the future of our technology. Identifying risk for scientists is such a complex issue, and there are so many aspects that make it incredibly challenging. I have the privilege of working extremely closely with the Science team at BenchSci to figure out solutions that empower scientists to increase the speed and efficacy of their research. It’s so rewarding, plus I get to work with machine learning and exercise my data engineering skills every day.

How have our machine learning capabilities grown in the last three years?

They’ve grown a lot! Our machine learning models are more sophisticated than ever, with much more advanced capabilities. Since I started at BenchSci, we’ve hired a lot of very talented engineers and established a solid, well-staffed Machine Learning Engineering team to look at new products and evaluate new models, as well as a Machine Learning Operations team to streamline the process. That, along with access to new technologies, has opened up so many possibilities for what we can accomplish to increase the value we offer our customers.

What’s the most difficult challenge of machine learning right now?

I think there are two aspects. One is the engineering behind it—that is, getting a system that works fluidly and is intuitive. Engineering those kinds of systems can be tricky, in part because the field of machine learning operations is still fairly new. The second challenge is understanding how to create good training data. This is a difficult process that involves both our data experts and machine learning engineers. The performance of the model is dependent on the quality of the dataset and the methodology behind it. I think that’s really cool because there’s no manual to tell you what to do or how to approach a problem—you need to figure it out yourself through trial and error. That’s especially true when it comes to developing models that work for the biological sciences.

You’ve participated in panels for women in AI and are a member of BenchShe+; what career advice do you wish someone had told you sooner as a woman working in data science, machine learning, and artificial intelligence?

I can be fairly straightforward when it comes to working, and I can also be somewhat stubborn on occasion. At some point in my career, I got extremely afraid people were going to just see me as difficult or standoffish. I think that caused me to hit a bit of a wall mid-career—I didn’t feel secure voicing my opinions, so I kept them to myself. Eventually, after chatting with a few people I trusted and checking out some recommended books, I learned to accept that people will think what they’re going to think. It might not always be what you want them to think, but that’s okay. You should always be true to yourself—what others think is their problem, not yours.

You’ve had a lot of hobbies, from crocheting to hydroponics to chicken farming, but what is your favorite hobby?

This might actually be surprising, but my favorite hobby is one of the more common ones: reading. I try to experience as much as I can in my life, but literature can allow you to go places and see things that you simply can’t in the real world. It can give you a little escape into a universe that doesn’t exist, where you can experience things you otherwise never would. Right now, I’m reading a collection of short sci-fi stories. One that I absolutely adore is called “Fandom for Robots,” about a sentient robot getting sucked into fandom for an anime. That’s something you can’t see in real life, at least not yet! It’s so cute and sweet, and I can relate because I really got into anime and the fandoms around them when I was growing up. 

We’re hiring! If you’re interested in building a great career in science or data engineering, check out our Careers page. And subscribe to our blog to stay up to date with all things BenchSci.

Written By:
Jasmine Alkin (she/her)