Machine Learning Engineer

About the role

Join a small and mighty team of engineers at Supernormal working on pulling the future of work closer to the present. Each engineer here is responsible for saving people thousands of hours of manual work every month, and we’re just getting warmed up.

As a machine learning engineer, you’ll focus on building and improving the AI stack powering Supernormal’s productivity suite. You’ll work closely with our product engineers and designers to evolve the AI along with the product.

What you’ll work on

Prompt engineering using state-of-the-art techniques to improve the core meeting assistant scenarios

Fine tuning, assessing, and deploying large language models

Developing and improving business and product metrics driven by the quality and cost of AI usage

Designing, training and deploying custom machine learning models to augment the AI stack, e.g. to manage transcript quality, compress text without losing semantic meaning, remove defects, incorporate semi-structured data

What we’ll expect of you

You have a strong working background in machine learning, NLP, and large language models

A collaborative and open outlook — we’re all about lifting each other up and getting better every day. You should be, too.

A willingness to get deep into a problem even when it seems impossible. You’ll always have support from the team.

Confidence operating with high agency. We’ll work together to decide what’s important, but we’d love for you to bring (and build!) your own ideas.

You’ll leave “it works on my machine” at the door. You know better than to trust computers.

You’ll come in willing to learn why things are the way they are, then suggest a better way.

You’ll understand that there’s no difference between “my code” and “their code.” It’s our code and we’re all responsible for it.

You’ll approach code review through a “how can the team level up?” lens — let’s all get better, together.

What you can expect from us

The team’s full support. We’re a friendly bunch and are happy to pair, talk through, or otherwise assist any time.

Honest and timely feedback. We’re all better when we can have candid conversations about what is and isn’t working.

A deep willingness to listen to your ideas: how can the codebase, our product, or team be better?

A respect for your time outside of work. We all work hard here, but we never forget to rest and have fun.

Interviewing

We know that interviewing is a deeply tiring and time consuming experience. To that end, we’d try to simplify our conversations as much as possible. Here’s what that looks like:

Intro: you’ll speak to a founder and the Head of AI to discuss the role, your experience, and what working at Supernormal is like

ML design: you’ll pair with an engineer for up to 45 minutes to tackle a fictional, but realistic ML problem. We’re most interested in understanding how you approach problems.

Close: you’ll meet with another member of the team or two to chat through culture, your concerns or questions, and your ideas for Supernormal’s future.

Sound like a fit?

We’d love to hear from you. We’re happy to receive applications at all skill levels, though please note that Supernormal is fully distributed. We’d gently suggest looking at other opportunities if remote work isn’t quite for you.