2. Journey into Machine Learning and Data Science Part II

Written by Salomey Osei.

The next part of the session was led by the second moderator, Salomey. This was an in-depth discussion of the challenges that they faced during their journey.

Question: Challenges during the journey and how you overcame them? How would you advise someone who is into ML?

Nyalleng: US has a culture where women are few in a class (course) and that women are not supposed to be in such a class. Many people ask questions that undermine you and make you question yourself, your worth, and your capabilities. For example questions like "Are you the new intern", Not to say that it is not good to be an intern but it almost looks as if you can only be an intern and nothing more even if you had a higher position than that. The question will always be "Are you an intern? The biggest question about these microaggressions is when you start to lose yourself. It is important to know your value so that you do not focus so much on proving yourself other than doing what you love to do. Those have been the biggest obstacle to my success. How I managed to overcome this is to go back to my core and remember why I joined the area in the first place, to also involve myself in communities of people I work with to share my knowledge and build myself without being belittled.

Usher: my challenges are a bit different as I am autistic. In India Autism is still considered taboo. When Autism is mentioned, what comes to mind is someone who cannot walk, talk or engage in any activity but needs a lot of help. So as an autistic person I focus on the positives, my brain filters a lot of things and I do not have to worry about filtering what goes into my mind. I was fortunate to be blessed with educated parents who understood these things and made sure they helped in any way that they can. My biggest challenge is the fact that autism was still considered taboo in India.

Amerley: Since I was coming from a Humanities and social background, I did not have enough background from the onset and did not know what to expect. When I wanted to help solve the problem of Ghana, I realized that I needed technology and tools. One of the challenges was "Introduction to Java programming" because of my background. I sat throughout the night to understand it. I decided that this is what I needed so I decided to calm down and face my fears. I reached out to friends from Ghana who also jumped into another challenge because at the time Ghana had very limited internet and the prices were paid per minute. I gradually bought books on programming and revisited techniques that my mother used to teach me fractions and with time, I got through it. After my studies, I came back to Ghana to work and realized that after all the hard skills and techniques, what kept me going was my soft skills which are being able to communicate with people and brainstorming as half of the time I tried to solve someone's problem. Also, In the early 2000's, data was a great challenge which we did not even find at the time. Currently, data seems to be available even for practice online.

Sarah Hooker: the first part is showing up for community events which gives you the platform to build yourself and to know more about people in the community. The most difficult part of starting this journey is to have courage and to have the courage to follow your curiosity. The common thing amongst all of us is that we were curious to solve a problem around us and in our communities but translating that and getting the tools for that s sometimes very scary. Early on in my technical journey, there were days that I could feel like I cannot do it and getting past it is the biggest challenge. The inception of ML is not more than 20 years so the idea of teaching and pushing yourself can be sometimes lonely so it's how you get past this. At this point, it is not just about your technical abilities, it is a brute-force effort but that will come. It will be like learning a new language and you will have to invest enough hours but you will be able to achieve if you invest the hours. The difficulty will be the mental strength to believe that it will come. What I have found in terms of concrete suggestions is to 1.) Narrow your vision because the truth is that it gets overwhelming when you believe that the task is learning everything. The best way to narrow your vision is to work on a problem because it narrows your scope and it makes your objective for the next 24 hrs of solving the problems. The biggest challenge as a newcomer is hopping from course to course and from resource to resource because it will be so overwhelming. So in my journey, I started using ML by working with nonprofit organizations in the communities I grew up in and trying to help them with data. I worked with volunteers who were engineers and researchers and I was learning so much from them. 2.) I will suggest working with people who are better than you particularly early on to accelerate your learning process because later on you are likely in a senior position and it is harder to get the same dynamics. 3.) The courage that Nyalleng talked about, after 7 years, if I am being honest is that there are big challenges. In the beginning, I buried it and thought that I will just get through the next 24 hrs. But the longer you get in tech the more you realize that it is not just about getting through the next 24 hrs but it's also changing the structure of how you work technically and being candid about what you are experiencing. We must talk about what we are experiencing because sometimes particularly women we experience things in isolation and we tend to blame ourselves, the truth is that communities like this are so important so we speak about the patterns we see so we can help each other. You are not going to make it alone, you are going to need someone who you can be yourself around and you can say to that someone that yesterday was awful and I barely made it and that you can trust them to keep it as that will give you the courage to show up.

Sarah Oppan: the challenges that the others have mentioned are similar to mine. How I tackle it is the following; 1.) You will find corporate politics everywhere. When ML is placed in Technology, then you find yourself in the men's world. I do not allow that to bother me, I allow my work to speak for me. Because when you go first, they ask similar questions to what Nyalleng said: "Are you an intern?" I use the work that I do to present to them what they do not know or what you have identified. 2.) The other challenge we see is when you are trying to communicate what you have done to executives where they have the perception of their business over a long period and you come in with a finding based on that same data where you have thrown in few parameters and it is changing what they believe in. Sometimes there is always this rejection and saying it is not right. They also argue about who are you because they have always had the same result over the years so what are you changing? 3) In the West, there is rich data. In our part of the world, the data you have is the data you know. So the data we have is unclean and few, it might mess up your work. In building my model, I try to make do with what I have and make sure that what I am doing is right.

Thank you for reading.

You can watch the full video here.

Contact/social media handle/website

Connect with us on Twitter

Join our Meetup Page

Visit our website

Email us at accra@wimlds.org