Lawrence Moruye’s data science journey is, in a lot of ways, the story of Zindi. He was in second year when he realised that data science was a promising career for someone like him, looking to combine a passion for mathematics and computer science. Just a few months later, Alfred Ongere of AI Kenya invited him to the Nairobi hackathon that launched the Zindi platform.
“At the time of the hackathon, I had some data science basics, but I felt that my skills were not good enough so I didn’t participate on the weekend,” says Lawrence. “But the competition stayed open for a few months after the hackathon, so I went away, sharpened my skills and made my first submission on Zindi.”
His model didn’t do very well — Lawrence placed 81st in that competition — but he was not discouraged. He read a feature engineering tutorial from Mohammed Jedidi (Zindi Ambassador for Tunisia), who won the challenge using a simple XGBoost, and was determined to do better next time.
What is Feature Engineering? A tutorial from Mohamed Salem Jedidi
"At the end of the day, some machine learning projects succeed and some fail. What makes the difference?
When the Mtoto News Childline Kenya Call Volume Prediction Challenge came around in 2019, he was ready. He’d practiced on a few other challenges in the meantime, and just completed a time-series course at university.
“I was fresh on the skills needed and decided to give this competition my best efforts. I read research papers on handling time-series challenges, and tried to combine that with my university knowledge. I was able to come up with a good solution — that was the first challenge I won on Zindi.”
Meet the winners of the Mtoto News and Childline Call Volume Prediction Challenge
Read more about Lawrence Moruye’s winning solution to predicting call volumes to Childline Kenya.
Since then, Lawrence has gone from strength to strength. He has placed in the top 3 of prized competitions several times, and now sits at 11th place on the Zindi leaderboard. He is Student Ambassador for Zindi at the Multimedia University of Kenya, and was a mentor in our recent Zindi Mentorship Programme.
“Lawrence was a wonderful mentor,” says John Godday, one of the junior data scientists that Lawrence mentored. “He dedicated his time, shared his knowledge and success stories, and helped us understand how to approach a range of data science problems problems.”
Zindi recently connected Lawrence with an internship opportunity at Qhala, where he is helping Mtoto News build an AI-powered mental health companion. This builds on some of the work he did in the challenge he won in 2019, as well as the more recent Basic Needs Basic Rights Kenya — Tech4MentalHealth Challenge.
“We are currently working on putting winning Zindi solutions into production, and building a chatbot that incorporates those models,” he explains. “It’s been really enjoyable, and I’m developing a great love for natural language processing (NLP).”
Getting to the next level
Lawrence’s advice for data scientists starting their journey with Zindi is simple: If you work hard and are willing to learn, you will go far. He also has advice on how to make the best use of the many data science courses available online.
“Doing a course on its own doesn’t always help, because the datasets they give you to use are very neat. Trying to combine things you’ve learned in different courses, trying to implement what you’ve learned in a specific challenge on Zindi, that’s the most valuable thing you can get out of data science courses. If I learn something but don’t put it into practice, it won’t be useful.”