Up until this very moment, nothing in the world has sparked my interest as much as Artificial Intelligence had, and for as long as I can remember, I’ve desperately wanted to plug myself into the journey of bridging the gap between human intelligence and computer intelligence.
2015 was the year I actively joined the GDG in my university, and being a huge game lover, I became really enthusiastic about the technologies we got exposed to during GDG events.
In no time, I had transitioned from being a volunteer to the lead facilitator for every event. I started leading codelabs, facilitating study jams, mentoring learners, and co-organising events for the 500+ developers and tech enthusiasts in our university.
Before long, we had gathered enormous success stories:
-
Our Google Play Store account had the highest number of apps in all of the GDG chapters and communities in Africa
-
Two developers (out of 5 winners across Africa) from our local community won the first Google-sponsored Udacity Nanodegree scholarships
-
We got featured in Google I/O 2016 as one of the most active GDG communities in Africa.
One very crucial move regarding my interest in AI turned out to be my final year thesis, which was based on Machine Learning for diagnosing cardiovascular diseases, and got published on the International Journal of Computer Science and Information Security (see paper). As an implementation of the research work, I published an Android application that acted as a chat bot for patients to enter their symptoms and perform basic diagnostic functions to help predict what possible heart condition it was.
Some months later, I had taken great interest in a number of Google Machine Learning APIs, and went ahead to build Toodoo — an ML empowered Firebase application that basically lets one add tasks to be reminded of, but with Visual Optical Character Recognition (recognizes text via camera), and Speech Recognition.
After graduation, I had applied and got the opportunity to lead a GDG Cloud community in Ado-Ekiti, where I continued to drive technological growth across Ekiti state focused on such events as Cloud Study Jams, Next Extended events, and TensorFlow Dev Summit Extended events, and have so far achieved the following:
-
Held over 17 meet-ups squarely focused on Google Cloud tooling knowledge-base for developer ease
-
Grown as a local developer community of over 750 developers and tech enthusiasts
-
Effected the birth of 5+ startups in Ekiti state, including Paey, DeliveryNow, Transporta, etc, whose products' stack are powered by GCP, Firebase, etc.
Sometime in 2018, shortly after I had begun my TensorFlow journey, I watched one of the Cloud Next ‘18 videos on youtube where the amazing Sara Robinson had introduced Google Cloud AutoML and its abilities. I immediately went ahead to try a small number of images of three different labels of lemonade bottles, and was really impressed with the results, which I shared in my talk during GDG Makurdi’s DevFest in 2018. And for a while, I pondered about how AutoML automatically calculates learning rate and decides the best calculations for cost/loss and optimisation functions, and be able to perform on the same level with neural networks designed by experts. I later found an article on Google AI blog about NasNet which explained that process to some extent. Cool stuff, yeah!
I have also had the opportunity to speak about topics ranging from Google Cloud ML APIs to Cloud AutoML to TensorFlow usages, etc. at several events including DevFest Nairobi, DevFest Warri, Switch Conference, DevFest Makurdi, etc. See some of my talks here.
I try to do a bit of writing (when I can) on a personal blog: coffeewithpeter.com. I am currently compiling a series titled “The Flow of Tensors”, and will begin publishing soon. TensorBots, let’s roll! :)
Earlier this year, I had resumed Lagos for a Data Wrangling role as a contractor, and a few months later, the general TensorFlow 2.0 Docs Sprint campaign began. And although I knew very little about organising events in Lagos, the excitement and success stories from the last TensorFlow Dev Summit in Ado-Ekiti still clung on. So, I reached out to the GDG Cloud Lagos team, and proposed to organise a TF Docs Sprint on behalf the GDG Cloud Lagos community, and yes, they accepted :)
The TensorFlow 2.0 Docs Sprint Lagos event, which was held on the first of June and had about 30 attendees, turned out to be a great inspiration for the start of even more intentional ML momentum in Lagos, including:
- Training continuity by the new TensorFlow Lagos community
Later on, I joined Semicolon Africa as an AI/ML Instructor, where my role includes leading them through understanding machine learning as an industry-relevant technological concept, teaching them what tools and techniques there are in ML (and why and what for), and guiding them through the process of applying ML in their various projects, as well as developing Curriculum for the same purpose.
Back in Ekiti State University, where I graduated from, wind had gotten to the department about my activities revolving around AI and ML, and the new Head of Department, Dr Mrs Oyinloye, had just recently introduced a new course titled Special Topics, which comprised of both Introduction to Mobile development and Introduction to Machine Learning. She reached out to me, and we had a conversation about my learnings around practical ML, after which she felt I could comfortably teach the ML part of the new course for a short period of time. Of course, I agreed :)
On Monday Sept 23rd, I resumed Ekiti State University for a short-term lecturing role teaching Introduction to Machine Learning as a mini-course. I am excited for them as the syllabus not only included basic ML with scikit-learn, but also how to build neural nets with TensorFlow.
It ends soon, and I get to resume back at Semicolon Africa.
For as long as I can remember, I’ve desperately wanted to plug myself into the journey of bridging the gap between human intelligence and computer intelligence.
And I have only just begun.