You’ve done it. You’ve nailed the interview, signed your employment agreement, and you’re officially ready for your first week as a machine learning engineer. Follow along with the steps in this handy guide to help build a strategy for how to excel in your first week on the job—and beyond.
Map the territory
As someone taking an interest in the future evolution of artificial intelligence and its rapidly changing influence on business enterprises, you are at the heart of a totally new way of doing things. This new, challenging field requires specific expertise in the evolution of practical computer science.
The overall purpose of a machine learning engineering position is to act as a link between the statistical and model-building work of the data scientist, and the building of production-ready, robust ML and AI systems, platforms, and services, in many cases collaborating with software engineering teams in the process.
Depending on the specific circumstances of your position, you may be assisting a team of data scientists to produce cutting-edge models for a given domain, and it will be your responsibility of translating said models into accessible builds for customers, business clients, government agencies, or even the public.
It’s important to keep in mind that your role is very intricately linked to many other key roles in any organization. Sébastien Arnaud, a Springboard mentor in the Machine Learning Engineering Career Track and a principal data scientist at Steppingblocks, notes: “Your first week should be spent connecting with everyone, and attempting to learn as much as possible about the current stack, process, the platform you will be deploying on, the current state of all the models in production, the ones still being incubated—and the role those models play in the various products of the company.”
Your day to day responsibilities could be a little different if you are working for a company that is developing ML/AI explicitly for internal use, but the goal remains the same: implement discrete statistical analysis/ML/AI into high-performance, high-availability production level systems that provide quick and easy access to all interested users or other system administrators.
Here is a shortlist of specific tasks that your role might include:
- Creation of web services/APIs for serving ML/AI model results and enabling access to customers or internal teams
- Automating model training and evaluation processes
- Translating the work of data scientists from environments such as Python/R notebooks analytics applications
- Automating engineering features, ensuring data for model training is clean and accessible to the team
- Facilitation of the flow of data between ML/AI models and an organization’s data systems
- Collecting data from data sources, blending them together, extracting KPIs, cleaning and standardizing the results
The machine learning engineering field is still in its early days of development, and relatively few programmers have all the acquired skills that are necessary to implement ML solutions in every case, but the industry is rapidly expanding. In your role as a machine learning engineer, you will often work alongside a diverse range of other programmers with different angles of expertise. Getting a grasp on all the new developments on the horizon will give you an edge.
Each year there are hundreds of machine learning and artificial intelligence conferences, ranging from small regional team collectives to must-attend industry-wide establishments. Taking an active approach to these events, and making an effort to continue bridging the gaps between the different areas of data science, will stimulate your capacity for crafting premium system solutions—and maybe even fire up your competitive side.
Tackle the tech
During this first week, it’s important to get comfortable with the flow of working with a team and contributing on a regular basis. The skills you learn during these small preliminary projects will help you in the future.
In your first handful of meetings, you are going to get a feel for your chief engineer and/or the reporting hierarchy of your company’s engineering and data science teams. It is likely that you will sit in for a series of presentations where the head of your department will outline the client or company that you are working for. Take note of the existing workflows used by the other analysts and engineers you will be working with.
Arnaud mentions that some of his most frequented tools are Jupyter Notebooks (in any flavor: Collab, DeepNotes) because they allow collaboration with everyone on the team. Additionally, Arnaud also recommends getting to know Streamlit, as it “offers a very easy way to build a small interactive app in a matter of hours and share it with your team.”
These tools, of course, will differ—so prepare some guided questions to speed up your integration:
- What kind of programming language do they prefer/expect? Do they use Python, R, Java, or C++?
- What type of frameworks does the team like to work with (ex. Hadoop, Spark, Pig, Hive, Flume?
- What ML libraries is the team familiar with? What do they recommend? (ex. Scikit learn, Theano, Tensorflow, Matplotlib, Caffe)
There are a whole host of libraries and frameworks at your disposal as an ML engineer. Be open to asking for assistance from your peers, reaching out to the greater community, and seeking new answers to new problems—but also don’t always try and reinvent the wheel. “Always stand on the shoulders of giants when possible,” Arnaud says. “Why reimplement a complex NLP Transformer model when other engineers have spent the time testing their implementation thoroughly?”
A good ML engineer can observe a task being performed by people and understand how to break it down in a way that would allow for automation, but they have to understand the capabilities of the technology first.
Communicate without computers (sometimes)
If this is your first ever professional position in the data sciences or AI research community, you might initially feel somewhat intimidated by your colleagues, hesitant to ask questions that might betray your lack of expertise. However, if you choose this route, you’ll be drowning in arxiv.org research papers and lost in Google search terms, up all night trying to retroactively engineer from outdated maps.
Keep in mind that machine learning engineering is a team sport. Arnaud says incoming engineers should expect to work with data engineers, data scientists, product managers, DevOps engineers, CloudOps engineers, testing engineers, and other software engineers.
Don’t let your instinct for pride get in the way of providing the best data services to your client or team or taking the opportunity to improve your skills. There’s no way a person can be good at everything. Sometimes what separates a good co-designer from a great one is the knowledge of when to consult another expert for backup. Because your work will be so interdependent, Arnaud notes that “making an extra effort to spend as much time as possible to understand what all the various areas of expertise are is key to your success.”
Oftentimes, what makes or breaks a team or obstructs its objective goals, is not a question of the programming or data, but the humans who are doing the work. Though your technical expertise will determine your capacity to effectively handle your workload, being an effective team engineer often requires additional soft skills. At the end of the day, don’t overlook the importance of the human element.
Here are some qualities that machine learning and AI engineers field look for in their co-engineers:
- Comfortable with failure. Frequently models will not work properly, or you will hit an unexpected snag translating data from analysis to modeling. Handling these with grace and patience can change the whole trajectory of a project.
- Driven by curiosity. The ML/AI field is booming at an incredibly rapid rate. The best people for the job have the passion and curiosity for the process, and for the potential of the algorithmic process.
- Putting yourself in the minds of the users/customers. Being open to the suggestions of others, and interpreting the world of feedback and data that you will come across (even in your first month), will be as important as the models you build
- Know when to stop. Data systems are iterative, meaning there is always and forever room for improvement on your projects. Time management is key and good engineers know when it is the right time to seal a project and deliver it to the client.
Find a mentor
Learning by yourself can get you off the ground, but it really takes the input of an experienced developer to level you up from a data science pre-career sandbox to industry-ready production code. Arnaud says: “One of the best ways to start building relationships is to listen and be inquisitive. Don’t hesitate to ask questions, and ask if someone on the team could take some time to walk you through an existing model, or be available for questions.”
Identify a senior member of your engineering team, or your department, and be open about seeking guidance on how to best effectively deploy yourself towards particular tasks. “Most team members are proud of the work they do and being genuinely interested in their work and their contribution helps break the ice,” continued Arnaud. “Don’t hesitate also to offer your help if you see areas when you could contribute.”
Coming into a community and a company (not to mention a whole company culture) with a fresh pair of eyes can be a huge advantage. However, it is crucial to recognize that as you adjust to your new role and your new team, there are going to be obstacles that will catch you by surprise. That said, always keep your mind on the process, and don’t get wrapped up in the hype machine and lose sight of the new and exciting work.
The field of machine learning and model programming is on the absolute cutting edge of human technological advancement—and you’re on the ground floor.
Is machine learning engineering the right career for you?
Knowing machine learning and deep learning concepts is important—but not enough to get you hired. According to hiring managers, most job seekers lack the engineering skills to perform the job. This is why more than 50% of Springboard’s Machine Learning Career Track curriculum is focused on production engineering skills. In this course, you’ll design a machine learning/deep learning system, build a prototype, and deploy a running application that can be accessed via API or web service. No other bootcamp does this.
Our machine learning training will teach you linear and logistical regression, anomaly detection, cleaning, and transforming data. We’ll also teach you the most in-demand ML models and algorithms you’ll need to know to succeed. For each model, you will learn how it works conceptually first, then the applied mathematics necessary to implement it, and finally learn to test and train them.
Find out if you’re eligible for Springboard’s Machine Learning Career Track.