Essa Mohamedali was the 3rd speaker in the Alliance4ai ChatGPT at work series, and he reminded us on why he is the most famous non-technical pioneer of the African AI ecosystem.
He showed how he used ChatGPT to place 62 out of hundreds in one of Zindi’s AI hackathons for datascientists! After which he shared about using ChatGPT for grant-writing, the actual purpose for which we had given him the mic. Here’s a breakdown of the tips he shared.
Coding like a rockstar with ChatGPT
Zindi is the premier platform for AI hackathons by Africa’s finest data professionals, and Essa decided to figure out if he could place amongst them. He signed up for a challenge, and described his process using ChatGPT Code interpreter as follows
He provided enough of the hackathon question to the tool
It started doing data analysis to tell him what’s missing, what trends are showing up and other insights.
Then it visualized the data to extract even further insights
Then it advised him on data preparation steps needed like cleaning, feature engineering and model building..
He asked it to build the model with its proposed algorithms and it went through logistic regression, random forest, and others until it came to a fair enough accuracy level.
Then he provided it the submission format for it to organize the final answer properly
All that was left was for him to upload the results to zindi, and voila, he was rank 62! His next goal is to aim for a top 10 finish, as this exercise is helping him understand previously complex processes involved in building AI solutions
Combining ChatGPT’s finesse and your flavor during grant writing
Many obstacles impede brilliant people from putting together grant applications. One of them is overcoming the intimidation created by the typical size of these applications. As a Innovation Associate at World Food Program who now reviews many grant applications, Essa shared how tools like ChatGPT could be combined with the unique flavor of the applicant to quickly produce quality applications.
He’d first give the LLM the RFP, his context or idea, the questions of the application and then ask it to help answer the questions like it was him. The weakness is that the answers it provides are more generic. So your job is to build your specificity into the large volume of text it helped you create.
It also has this strength to alter your content to fit the tone and style that the application requests, making you better align with the grant provider and closer to your goals.
Now the ball’s in your court: what project will you work on next using ChatGPT?
About Alliance4ai’s career growth mission
To inspire Africa’s Youth to be constantly aware of Technology trends and to apply them at work in ways that are responsible and improve society, we brought together volunteers to create the Alliance4ai WebApp where people signup to Circles in their countries or universities and have engaged discussions about key topics bi-weekly. Invite people in your network to join, and let’s build the Africa we want!