machine learning
machine learning

Year

2022

Category

Web & Mobile

Year

2022

Category

Web & Mobile

Making Machine Learning Accessible for Beginners and Educators

Zelda is an educational deep learning framework designed to make machine learning (ML) accessible to beginners and educators. I was responsible for designing and implementing an intuitive web platform using Tilda, creating a brand style, and developing custom illustrations that simplify technical concepts.

Problem

Machine learning frameworks often overwhelm newcomers with:

  • Dense, technical documentation

  • Interfaces designed for advanced users

  • Lack of approachable teaching tools

Zelda needed a platform that made ML concepts easy to understand, visually engaging, and supportive for both learners and educators.

Challenges

  • No-code constraints: Working with Tilda meant limited customization and interactivity.

  • No developer support: Required creative solutions to deliver complex features without code.

  • Diverse audience: Needed to balance clarity for beginners with enough depth for educators.

Research

User Interviews

  • Beginners struggling with jargon and complex onboarding

  • Educators seeking structured resources for teaching

Personas & Flows

  • Developed personas and mapped user flows to define intuitive pathways to tutorials, documentation, and installation guides.

Competitive Analysis

  • Benchmarked against TensorFlow and PyTorch, identifying opportunities for clearer navigation and friendlier onboarding.

Feature Prioritization

  • I used MoSCoW method to prioritize quick-start guides, code examples, and interactive tutorials while deferring advanced tools unnecessary for the audience.

Outcome

I delivered a professional, scalable, and fully responsive web platform that empowers users with interactive tutorials, streamlined learning paths for faster onboarding, and a vibrant community hub for Q&A and collaboration.

Visual Style

  • Created a brand style inspired by The Legend of Zelda series, incorporating a green-gold palette, soft gradients, and approachable typography to make learning feel less intimidating.

  • Designed a cohesive illustration system featuring icons, diagrams, and thematic artwork to explain technical processes visually.

Interactivity & Navigation

  • Streamlined navigation for seamless access to tutorials, documentation, and installation steps.

  • Designed clear CTAs placed strategically based on A/B testing results.

Key Learnings

  • Designing for accessibility in technical education requires visual storytelling and progressive complexity.

  • No-code platforms can deliver high-impact experiences when combined with strategic UX and brand consistency.

  • Custom illustrations are powerful in breaking down complex ML concepts for a broader audience.