5 Career Tips from Women Leaders in Machine Learning
Understanding how important representation, role models, and mentoring had been to my own career journey, I started a network to support other Amazon employees looking to pursue a career in machine learning (ML) and artificial intelligence (AI).
Open to anyone working at Amazon, the global Women in ML/AI group hosts regular networking events and organizes panel discussions with industry experts on career development.
To discuss learnings from our professional journey, I sat down with fellow board members, including senior documentation manager Michelle Luna, senior software development manager Anna Khabibullina and general manager and product lead Shubha Pant. Here are some of the advice we found invaluable when launching and building a career in the field.
1. Put Yourself out There and Make Connections
Luna, Khabibullina, Pant and I are all proof that there are many paths into ML and AI — from the traditional and linear, to the more unconventional.
I started out in the technology and media communications sector in Germany, where one of my first roles was in market research. This is where I realized that I wanted to understand the fundamentals of data science and ML. I have a business background, but I just kept building my network with people in the field and pursuing data science roles and internships.
Luna: “I had no real machine learning experience before I joined AWS. I had worked in language translation software 25 years ago, so I was sort of pulling at a thread from a past career, but this experience seemed to get me in the door. I have some DevOps experience too, and this applies to my role now in ML, which I hadn’t even realized.
I would say don’t be afraid of putting yourself out there, no matter what your career path in technology has been. One big thing our members want is a place to network with other women who already work in ML. I can’t emphasize enough the importance of reaching out and building those connections.”
Khabibullina: “I joined Amazon as a traditional software development engineer in Amazon Robotics. When I got curious about ML, I reached out to a few teams in that space, including the Alexa team, which I joined shortly after.
This was my first experience of ML at Amazon, and I absolutely loved it. In particular, I thought the smart home use cases — from improving home security, to enabling graceful aging — were so powerful.
Since then, I’ve been dedicated to building ML-based applications for the last five years and counting. The door might seem closed, but sometimes a light push is all it takes to open it.”
2. Use Your Background as Your Unique Selling Point
It was interesting how all of us found out that people interested in ML/AI will be surprised at how much they already know.
Luna: “I always say to women in our group, you probably know more than you think you do. Any background in computing is relevant to ML and AI. Most specialties across a range of areas in computing have started moving into ML, and that background will help you more than you realize.”
Khabibullina: “The ML space is developing so rapidly. Four years ago, the most common roles in ML that I could see were scientist and engineer only. Today, we need more specialized backgrounds to cover emerging needs, such as data-specific roles, applied research scientist roles, machine learning engineer roles, DevOps roles, frontend-, cloud- and software-engineer roles and more. As the field continues to expand, people with diverse backgrounds are essential.”
As for myself, before becoming an AWS data scientist, I was in another role at Amazon. I wasn’t always confident that I could use my interest in ML/AI to really identify as a data scientist. But a mentor directed me to websites and other resources, and often asked if I could understand the content. When I replied that I could, she told me that, yes, I was a data scientist and to go for it. She really validated my ambitions and encouraged me to use my business background as my unique selling point.
3. Explore Online Courses to Help You Get Started
One big perceived barrier to ML and AI is the idea that you need to be highly qualified and accredited to work in this space.
Pant: “I meet many people who think that they have to have a Ph.D. in statistics or math to start any role in AI/ML, which is just not true. It’s a total myth that you must go through a lot of courses and get a specific degree. You can get started without this. Any AI/ML organization has typical roles like product manager, technical program management and data analyst, which all require domain expertise but not a deep AI/ML background.”
Luna: “As far as skill sets are concerned, transferable skills like communication and attention to detail will take you far in this field. If you’re really interested in launching an ML career, I would build an understanding of basic frameworks such as Python, Java or Node.js. Understanding just one of these can be very helpful. The industry sits at a cross-section of a lot of different skill sets, so design and customer experience are super valuable too.”
In my case, when I knew I wanted to get into ML/AI, I decided I would do some further study. With my background in business, I just didn’t feel like I could do a master’s in mathematics. Instead, I found a course called “Big Data and Business Analytics,” which I studied for a year. It focused on the perspective of the customer, which I loved, and this has stood me well in launching my career as a data scientist.
4. Be Intentional When Choosing a Mentor
We agreed that a carefully chosen mentor can change your career, and it’s also a great way to give back once you break into the industry.
Khabibullina: “For me personally, mentors have made a dramatic impact on my career and my ability to grow. I try to learn one thing from each mentor, be it a soft skill such as how to influence without authority, or a technical one, such as the best approach to developing an ML model for a specific use case. When working in ML/AI, you should treat finding a mentor in the same way scientists use research and multiple perspectives from leading scientists, to help them forge a path forward. By this I mean being intentional about who you ask, considering their skill set and knowledge, and thinking about how you can help each other.”
Luna: “In general, as a woman in computing, it’s really important to have a mentor, and in ML in particular it will help you uncover exciting new areas you might not yet know about. This is a big focus in our group. We mentor in informal and formal ways to help women connect and share, including how to build career development plans for this industry.”
“We have a Slack channel for the group, and it’s great to see our members put themselves out there. Someone will ask for a mentor who specializes in something specific, then someone else will step forward, and they arrange to have a coffee chat. More formally, we have what we call mentoring circles, which are four circles of small group mentoring that enable women to share stories and listen to each other in a more intimate setting.”
5. Look for Unmet Needs
ML and AI is a relatively young field. With so many opportunities and so far to go, there’s a chance for more women to get a seat at the table.
Luna: “There’s this myth that ML and AI is a narrow space where we work on voice commands and speech recognition, but it’s so unbelievably broad. In the last six months alone, my team has launched ML services across DevOps, computer vision, manufacturing and healthcare. There is just so much happening and so many opportunities to work in an area you’re interested in.”
Khabibullina: “This is an industry on the edge of innovation, and sci-fi movies might misinform us that it is far more advanced than we think. The truth is, it’s still very early days, and the opportunity is here to set the standard in diversity and representation. ML represents our chance to set expectations in the technology sector, to build out a field without any prior bias as to who should be working here.”
“Beyond representation, individuals have an opportunity to design their own career path. The field isn’t saturated yet. We need more people and more expertise. It means you can decide where to go and what to work on, and achieve this more easily than you think, instead of competing for jobs in a busy market.”
Pant: “I don’t think it’s possible for anyone to imagine how far ML can go as an industry. This is why I tell people to never accept that anything is out of bounds. Always look for unsolved problems and unmet needs. When you take the ownership, believe me, the solution will follow.”
“All you have to do to stand out and make your mark is to look for unmet needs and solve challenges others don’t want to take on. Initially, this will be uncomfortable, as things can be ambiguous and vague, but it’s unexplored territory. There’s a lot of growth and success there for the taking, especially for women.”
Making Machine Learning More Accessible
AWS recently announced two new initiatives designed to open educational opportunities to people interested in learning about, and experimenting with, machine learning technology.
The AWS AI & ML Scholarship is a $10 million education and scholarship program specifically aimed at helping underrepresented and underserved high school and college students from all over the world learn foundational ML concepts and prepare for careers in AI and ML. Delivered in collaboration with Intel and talent transformation platform Udacity, it will offer students free access to dozens of hours of training modules and tutorials on the basics of ML and its real-world applications, a chance to win a Udacity Nanodegree program scholarship and mentorship opportunities. Find out how to get started.
AWS is also offering free access to a version of Amazon Sagemaker — a service used by developers, data scientists and researchers worldwide to build, train and deploy ML models quickly — through the new Amazon SageMaker Studio Lab. The lab, which doesn’t require users to have an AWS account or provide billing details to access it, will enable people to start work on ML projects in the time it takes to open a laptop. Users simply sign up with their email address through a web browser and can start building and training ML models with no financial obligation or long-term commitment. Learn more about Amazon SageMaker Studio Lab.