We are always winners: Mohamed Salem Jedidi talks data science competitions and being a Zindi ambassador
“I joined Zindi and became an ambassador because I loved the idea that we would have an African platform for data science competition. My main goal was to build and support the African community of data science.”
Mohamed Salem Jedidi is one of Zindi’s first and most accomplished users, and part of our first group of ambassadors. He consistently performs well in Zindi challenges and has won several, including the Traffic Jam challenge, Zindi’s largest prized competition to date. He is currently ranked #2 on Zindi’s leaderboard.
“Data science competitions are like a race, and you should be proud of yourself when you win, especially when it is an international competition because you are representing your country and your community. But actually, we are always winners: if we don’t win a prize, we win in new knowledge.”
From novice to data science master
This commitment to knowledge and learning has taken Mohamed from a Kaggle novice to one of InstaDeep’s senior data scientists in just a few years. These days he uses machine and deep learning models to build multipurpose AI agents solving problems in a wide range of industries, but in 2017 he was new to machine learning.
“In the summer of 2017 I discovered the Kaggle platform; it was like magic for me and I knew immediately, this is where I belong,” he says. “My first competition was the New York City Taxi Trip Duration, and I started with no ML knowledge at all.”
Mohamed believes the only way to learn is to practice, so he dived into books, articles, and blogs as well as analysing competitor kernels and notebooks, and tried to transform knowledge into practice. Two months of hard work later, his team was in 16th position — the top 2% of the leaderboard.
“This was a huge success for me. It was great motivation to carry on in the field, to continue playing in many different Kaggle challenges and learn more about ML. In each competition I won new knowledge and new techniques, and I met new people that I learned a lot from.”
Bringing expertise back to Africa
Joining Zindi in 2018 was a chance for Mohamed to start giving back to the community, and he has been a mentor and role model for a whole new generation of data scientists on the platform. He has some great advice for upcoming competitive data scientists and those new to Zindi:
“To get to the top of the leaderboard, you have to work smart and work hard. I suggest following competition discussions, and asking questions in the discussion forums, as this will help you to understand the task better. Look to a similar challenge and check the best solutions — this could help you to develop new ideas that could help you. Lastly, don’t close yourself off and rely on your own knowledge alone — this will not help you to learn new things, and you will never win or grow that way.”
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? Easily the…
But Mohamed is more than a skilled data science competitor. Since becoming ambassador for Zindi in Tunisia, he has helped to grow a long-standing partnership between Zindi and InstaDeep, and has helped organise exciting events like the AI Hack Tunisia 2019.
“Being a Zindi ambassador has allowed me to increase my professional network and meet new people across Africa,” he says. “It’s so amazing that I can know and be friends with at least one person from each country in Africa.”
Mohamed also enjoys working directly with the team at ZIndi, and provided technical support for our competitions throughout 2020.
“At a small startup, there is no shortage of work to do,” says Amy Bray, competitions lead at Zindi. “Mohamed’s expertise and knowledge of data science competitions has helped us in reviewing winner’s code and preparing data for competitions.”
In 2021, Mohamed hopes to see data scientists meeting up in person again.
“Let’s hope that 2021 is much better than 2020, and we can gather to meet each other again. I am really excited to help organise the Deep Learning Indaba in Tunisia, since this event was canceled last year due to coronavirus. I am also looking to grow the ML community in Tunisia by organising a series of events — I really hope we can do this event in real life and not online.”