From Code Newbie to GenAI Developer: My Journey Through the AI Capstone Project

When CreateHer Fest shared the opportunity to dive into generative AI, I jumped at the chance! Despite being familiar with Jupyter notebooks, building a few models (here’s to you SheCodes), and working with them professionally from previous data exploration projects, stepping into the Kaggle environment was a new frontier for me. Little did I know how transformative this journey would become.

What made this experience unique was the multi-faceted learning approach. I found myself:

• Reading whitepapers that once seemed impenetrable but gradually became comprehensible

• Listening to AI-generated podcast summaries that distilled complex concepts into digestible insights

• Attending live workshops with industry experts who shared real-world applications and best practices

Each learning method built upon the others, creating a rich tapestry of understanding that transformed theoretical concepts into practical knowledge. The combination of text, audio, and visual learning catered to different aspects of my learning style, making difficult concepts stick in ways they never had before.

My Capstone Project: Smart Task Parser

For my capstone, I created a Smart Task Parser that leverages three key generative AI capabilities:

1. Structured output - Converting natural language task descriptions into standardized JSON format

2. Few-shot prompting - Teaching the model how to interpret tasks through specific examples

3. Function calling - Implementing a system to fill predefined data structures based on user input

The system takes everyday phrases like "Buy cheese tomorrow" and transforms them into structured task data with categories, priorities, due dates, and time estimates. It's the kind of tool I wish I had in my daily workflow!

Debugging Adventures

The path wasn't without obstacles! I encountered a series of frustrating but educational challenges:

• Battling stubborn indentation errors that reminded me just how important consistent formatting is in Python

• Discovering the joy of "return outside function" errors when my code structure went awry

• Learning about API rate limits the hard way when my Gemini quota was suddenly exhausted

Each error became a learning opportunity, teaching me to debug methodically and adapt to unexpected constraints. When my API quota was depleted after just one successful task parse, I pivoted to simulated results to complete the visualization component, a real-world reminder that developers must always be ready with Plan B!

Despite the challenges, seeing the final visualizations come to life was incredibly rewarding. The bar charts showing task priorities and time estimates, alongside the pie chart of task categories, transformed abstract code into tangible insights.

The experience reinforced that the most valuable learning happens not when everything works perfectly, but when you face challenges that force you to dig deeper, research solutions, and implement creative workarounds.

Looking Forward

This capstone project marks not an endpoint but a beginning. The skills I've developed, from understanding generative AI capabilities to deploying them in practical applications, have opened doors I didn't even know existed a few months ago.

To anyone starting their AI journey: embrace the errors, celebrate the small victories, and know that each debugging session brings you closer to mastery. The field of generative AI is evolving rapidly, and there's never been a more exciting time to be part of this revolution!

Thank you, CreateHer Fest, for this incredible opportunity to grow, learn, and create. The journey continues!

Want to check out my Capstone? You can do so here.