What I Learned from Volunteering as a Machine Learning Teacher this Year
From September to December this year, I volunteered as a machine learning teacher for the ML & AI course at ReDI, a non-profit organization offering online and in-person programs. It was a unique and unexpected experience, so during this reflective period between Christmas and New Year, I took some time to reflect on how it changed my perspective on ML topics and helped me regain some "nerd spark":
What is ReDI?
First, if you've never heard of ReDI, here's what they say about themselves:
ReDI School of Digital Integration is a non-profit tech school providing access to free digital education. We speed-up job market integration for Newcomers and Locals alike!
Since 2016, we offer a variety of courses; from Computer Basics to more advanced tech courses. Additionally, we offer a unique career program including mentorship, career workshops, company visits, job matching and much more. A semester takes 3-4 months (part time) and the teachers are volunteer tech experts.
(Quote from: https://www.redi-school.org/our-impact)
This is the course I joined as part of the teaching team: https://www.redi-school.org/data-analytics/hamburg/dcp/machine-learning.
I was part of a team of eight volunteer teachers - professionals and graduate students from across Germany - who volunteered their time and knowledge throughout the 14-week semester to teach online, prepare lectures, grade homework, and answer questions during class.
Lesson 1: Teaching = Learning
A lot of these lessons are things I realized earlier in life but have since forgotten as life got busy and I joined the busy world of office workers. This is one of them.
I was a teaching assistant during my time at university (I have degrees in mathematics and computer science), and before that, I also gave tutoring lessons at school to classmates and younger students.
I've always loved teaching - this rewarding moment, when you break down a complex concept just right so the student understands, is unmatched. Not only do I help others understand - but each time I explain a concept, I revisit prerequisites I take for granted to build a compelling lecture or explanation. This process always leads to a deeper understanding for myself.
So, that's Lesson 1.1: there is no better way to test your understanding than to explain it to someone else.
Lesson 1.2 is also true: I learned so much after university from work experience - much of it I only realized I had learned when I taught this in the course.
A: Well, we certainly don't simply replace it with the mean. First, we ask the data owners or business experts, as there often is a reason for the missing data that can be encoded or used as a filter. Many times, they even point us to another data table that actually contains the missing data, solving the issue.
There were dozens of such interactions - teaching made me realize everything I learned passively, which was a very welcome realization.
Lesson 2: There are many basics without proper resources online
While some materials were provided by the ReDI school, the machine learning course is still very new, so we were strongly encouraged to rebuild the presentations and lecture materials. It's surprisingly difficult to cover "everything" about unsupervised learning in one 2h-lecture. Partly, that's probably because I'm a massive perfectionist (that's why you rarely read or hear from me).
But when looking for materials online to help me, I realized that it's still very hard to find quality material about ML basics online. Many materials are surface-level and don't explain any of the logic or mathematics behind the concepts. Often, they don't address pitfalls like applying a model to out-of-distribution data or other mistakes made by choosing your training and testing datasets leading to overconfident model evaluations. I don't need another "toy" notebook, I need actual advice on what to consider when applying the concepts to real life problems and datasets...
This has definitely encouraged me to try writing my own tutorials again - even if they are only used for the upcoming ReDI semesters and to be able to quickly send a link to someone with a question of material, where I know I agree with it.
Lessons 3: Communities like this are rare (and inspiring)
In daily life, people are busy with all sorts of problems in their own lives - so connecting with others isn’t always a top priority in daily life.
Being part of a community dedicated solely to learning and helping others was incredibly uplifting. Personally, academic pursuits align strongly with my personal values. I love learning and naturally vibe with others who are curious and interested in exploring new ideas.
Additionally, being part of a team of teachers - who also made time in their busy lives to teach for free - gave me back some ... spark, that I've lost somewhere in the last few years.
Many in society view a job as a necessary evil, dreaming of reducing hours, early retirement (right after starting their career), or leaving work early. That's not why I chose this career. I genuinely love mathematics and abstract concepts. I find it exciting to figure out how to apply these ideas through machine learning to create real-world value. Of course, I sometimes wish I could just take a nap or go on vacation, but there are never enough hours in the week to to work on all the cool projects that I want to explore and learn from.
Being genuinely excited about "work" (really, I'm just a huge nerd who happens to get paid for solving puzzles and challenges) made me feel like a weirdo in office environments. Meeting others who share my enthusiasm for machine learning and the craft has been incredibly healing.
Some negatives to keep it real
I don't want to give the impression that teaching was all sunshine and rainbows (although I will 100% teach again next semester starting in March if they're willing to have me). It was a lot of work, often underestimated - I’ve learned that reworking a two-hour lecture from scratch takes more than just an afternoon... Or at least, I hope I learned that lesson now 😅
And, not every participant in the class will share the same enthusiasm or priorities - we can't control that.
A specific complaint I've encountered is that some people seem to think volunteering online "doesn't count." On several occasions, I’ve mentioned I was unavailable that evening or simply shared that I was volunteering that day, only to receive dismissive reactions once it became clear I was teaching online.
Providing free education to individuals who might not otherwise have access to this knowledge or the career opportunities it offers feels like a meaningful way to improve lives in the long run. I recognize that I was very lucky to grow up with free access to graduate-level education in a relatively wealthy country and family, and I feel strongly about passing on this knowledge and privilege, even if only in small ways.
If you know someone in the field of IT in Germany, Sweden or Denmark, who might be interested in teaching or mentoring students, feel free to share this link: https://www.redi-school.org/become-a-volunteer-at-redi-school
I'm looking forward to hopefully another semester and sharing some lessons and tutorials here as well 😊
No spam, no sharing to third party. Only you and me.
Member discussion