Photo by Tara Winstead: https://www.pexels.com/photo/an-artificial-intelligence-illustration-on-the-wall-8849295/
AI is here to stay. Product managers must understand the value AI can bring to a product, the various options available, and how AI can be integrated into their products. Hackathons are a great way to gain this knowledge.
In this article, I will discuss my experience building AI-powered products for two recent hackathons.
Why hackathons?
I have been exploring ways to gain valuable experience in AI. There are loads of articles and courses available online. I enrolled in one of them on AI product management. While the course didn’t turn out to be as valuable as I expected, one of the key takeaways was the introduction to Devpost and the idea that I could enroll for a hackathon as a product manager.
Inspired by this idea, I enrolled in a hackathon as a product manager and looked for a group to team up with. Since I couldn’t join a team, I decided to build the product independently from start to finish. I learned so much from the hands-on experience in my first hackathon that I decided to follow the same approach for future ones.
How did I identify the problem I am trying to solve and how to solve it?
The first hackathon I enrolled in was Google’s Women Techmakers presents She Builds AI. The goal was to develop an AI-powered solution to address gender equity challenges.
Since this was my first hackathon and I didn’t have a team, I was completely clueless about where/how to start. I started brainstorming ideas and solutions with ChatGPT and the interactions proved extremely valuable.
ChatGPT provided examples of ideas and solutions that were helpful in understanding the general expectations of these hackathons. ChatGPT suggested some generic project ideas. I expanded on them by incorporating insights and knowledge from recent readings.
 |
|---|
| Photo by Blue Arauz: https://www.pexels.com/photo/wooden-framed-sign-with-friendly-message-29562925/ |
I followed a similar approach for my second hackathon, the Google Chrome Built-in AI Challenge. The goal of this hackathon was to develop a web application or Chrome Extension that uses one or more Chrome built-in AI APIs.
How did I build the products?
I spent much time exploring and selecting the right tools and technology for the projects. It took a lot of time to configure the tools, for testing and troubleshooting and pivoting to explore alternatives when I hit a dead-end.
Below is the final list of tools I used for each hackathon.
First hackathon:
- Data collection - used an Excel sheet
- Model fine-tuning - Google AI Studio with Gemini 1.5 Flash model
- Interface development - Thunkable
- API integration - accessed Gemini API using Thunkable
Second hackathon:
- App development: Built on Replit using Replit Agent.
- Content generation: Motivational quotes and reflective prompts are generated via the Prompt API from Built-AI for Chrome.
What challenges did I face?
Being a product manager with limited coding experience, the biggest challenge was the steep learning curve in mastering the tools that were used. Some of the tools were especially challenging to use as they were not intuitive and the documentation/help resources were not adequate for a beginner. The important part has been to recognize when to pivot to other tools when I hit a dead-end.
What were the biggest takeaways/learning?
- I learned to use various AI services, no-code platforms, and an online integrated development environment (IDE).
- I was pleasantly surprised to see the breadth of AI tools and APIs available and how easily one can access and integrate AI into products.
- The overall experience and hands-on experience have improved my technical knowledge and boosted my confidence in collaborating effectively with engineering teams.
You may see my submissions and demo of the products here and here.
I hope this blog encourages product managers with even miminal coding background to create AI-powered products. Thank you.