Enter your email and we'll share the insights. Hitting submit opts you into our mailing list.
A Machine Learning Engineer develops algorithms and models that enable machines to learn and make predictions or decisions without explicit instructions.
A Machine Learning Engineer trains a model to either predict or classify data that a normal statistics approach couldn’t handle. For example - let’s say an airline asks for customer feedback on Twitter and receives 10,000+ tweets back. Machine Learning Engineers can train a model to determine whether the feedback was good, neutral or negative. It’s a good way to get customer feedback into data.
Machine Learning and Deep Learning have both received quite a deal of press over the last several years, so it’s not a bad idea to understand the difference between them when hiring to fill these niche roles.
Machine learning has two different types of data sets - unsupervised learning and supervised learning. Unsupervised learning means you have data you don’t know how to label (one example of this could be stocks - you won’t know if a particular stock is profitable or unprofitable). Supervised learning means you have a data set that you know the labels for is for the medical field (one example of this could be you have a data set with confirmed individuals who have diabetes, and you’re looking at different data points that correspond to the person with diabetes). Machine learning works with both unsupervised and supervised data sets to develop models that can make predictions and accurately account for different events.
Unlike Machine Learning, in Deep Learning, your data sets won’t have any labels on them, and it will be hard for a human to label the data. Machine Learning Engineers working in Deep Learning will have to train a model to find patterns in unlabeled data, with tens of thousands of data points.
1. Basic coding and software development experience. At a minimum, Machine Learning Engineers should have experience in basic programming languages like Python and web applications like AWS.
2. Specialized knowledge in Machine Learning related technologies. Look for resume keywords like Pytorch, TensorFlow and PySpark.
3. Experience in building frameworks and algorithms themselves. Can your candidate quantify their accomplishments? For example, can they explain how their machine learning model improved search results or led to higher customer satisfaction? Dig into the specifics of their projects.
4. Involvement in projects, competitions, or organizations outside of a full-time role. Data competitions (like those on Kaggle) are popular and participation can be a strong indicator the candidate enjoys working in the field. We’ve also met Machine Learning Engineers who have created neural networks that analyze real estate prices, or have used Google Maps business data from to analyze data and build methodologies. Being involved in peer reviews of source code or displaying projects on Github is a positive indicator of a committed Engineer. Many large data sets are now publicly accessible, and savvy Machine Learning Engineers will know how to use them to pursue projects of interest on the side.
Many Machine Learning Engineers will have a background in computer science, coding or math from undergrad, or a Masters degree and will have studied Machine Learning itself. Many team leads who are directing the day to day work of Machine Learning Engineers will have advanced degrees, like PhDs, in Machine Learning.
Python, C++ or C sharp and MatLab and natural language processing can all be important for a Machine Learning Engineer.
1. Determine who the role will report into. Will this role report into a Director of Engineering? VP of Engineering? Your company’s Chief Technology Officer? Deciding who the role will work with can help you determine the level at which to advertise. Hiring for an individual contributor is different from a leadership role – from the job description to interview questions.
2. Write out the main project or projects you will want this individual to work on. Will they need to develop a specific machine learning application – for example, to create an algorithm to track and effectively sell to visitors on your website? Will they create new smart product features, or improve upon an existing product?
3. Write out your hiring process and decide if you want to include an assessment. Jot down how many rounds of interviews you want your candidates to complete, which important stakeholders they should meet, and whether a technical assessment will be helpful.
A typical day for a Machine Learning Engineer could be spent working on their model for a few hours and then meeting with their team to discuss projects. In the afternoon, they may spend some time troubleshooting a potential bug and then working with their team on some data manipulation.
When putting together interview questions, feel free to pull from our suggestions below. Here are some of the top interview questions we suggest to ask Machine Learning Engineer candidates.
1. Can you tell me about a time when you had to use large data sets to draw insights?
2. What’s an example of an algorithm you’ve created or worked on, and what did you design the algorithm to do?
3. What’s your experience with summarizing, extracting, or diarizing data?
4. What Machine Learning tools have you used, and which are you the most familiar with or expert in?
5. How do you clean or test code?
6. What’s your experience in developing user (or customer) facing systems?
7. What’s your favorite Machine Learning project you’ve worked on? Why was this your favorite?
We’ve recruited for many different Machine Learning roles, including:
Interested in learning more about Machine Learning Engineers, and how Recruiting from Scratch can help? Send us a note at sales@recruitingfromscratch.com.
Let our team help you get where you need to be.
