From Infra Engineer to ML Engineer: My Journey to Lyft

by phantom_flare749
Lyft
senior
Prep: 6 months
13
155

The past six months have been a whirlwind, and not just because I decided to pivot my career toward machine learning. With nearly seven years under my belt as a senior infra engineer, I had seen countless systems come together. Yet, my thirst for growth led me to the world of ML, where data and algorithms collide. It was time to level up.

The preparation phase was intense. I started by brushing up on the necessary ML concepts. That meant diving deep into algorithms, frameworks like TensorFlow and PyTorch, and brushing up on stats. My evenings were consumed by courses on platforms like Coursera, which helped me grasp the breadth of the subject. I mocked up projects using Kaggle datasets, trying to recreate models that I could proudly show. Each mini-project brought both triumphs and a few failures, including my infamous attempt to build a predictive model that performed worse than random guessing. Good fun.

As the interview date approached, the pressure kicked in. Seven rounds of interviews were looming ahead of me, so I decided to simulate the experience by organizing mock interviews. Some were with friends from my old work, while others were with strangers I paid via platforms for coding practice. Those sessions were invaluable, exposing me to unexpected questions and keeping the jitters at bay. However, one specific moment stands out: the time I totally blanked on implementing K-means clustering during a mock session. I sweat through that entire discussion, realizing I had to solidify my grasp on foundational algorithms even further.

When the real interviews began, it was a mixed bag. There were coding challenges, case studies, and behavioral questions about handling uncertainty and failure. I aced the first round of coding, thanks to my previous experience with system design. But things took a turn when I encountered a live coding session focused on neural networks. I couldn't recall the nuances of dropout layers and ended up fumbling through it.

The technical interviewers were keen. They didn't just want to see if I could write code; they were interested in how I framed problems and approached them. I finally found my stride in explaining my past projects and how I architected reliable systems, using my infra background as a unique angle. By the last round, I was exhausted but cautiously optimistic. My interviewer mentioned they were impressed with my analytical mindset, which I took as a good sign.

A week later, I received the call: they were extending an offer! I remember trying to contain my excitement as I hung up. Securing this role as an ML engineer at Lyft feels surreal, especially considering the journey it took to get here. I had danced with my doubts and triumphed with my curiosity.

It was about confronting the uncomfortable; facing the stress of a completely new domain, realizing that it’s okay to mess up and stumble through. With every challenging moment, I learned something new, and that made all the difference. And now? I can’t wait to dive into machine learning on a grand scale, with a company as innovative as Lyft.


Tips

Spend a significant amount of time on the foundational concepts of ML, don’t skip the basics., Conduct mock interviews with a mix of friends and strangers for varied feedback., Engage with open datasets and create projects that showcase your analytical abilities., Be humble about your limitations; it’s okay to admit when you don’t know something., Track your learning process and celebrate small wins to keep motivation high.


Markdown supported