STYLA

Framing the Problem

What is the problem you are trying to solve? Who does it affect?

Choosing what to wear each day gets repetitive and time consuming.

Most people struggle with:

• Planning outfits tailored to their body, occasion, and weather.

• Making the most of their existing wardrobe. People tend to forget what they have or default to the same outfits all the time.

• Knowing what to buy that matches their style and fits.

• Getting styling advice that's personal, on-demand, and actually useful.

Idea Explanation

What is your idea? How does it fix the problem?

We propose a digital personal stylist that provides you with outfits based on specific events in your calendar.

Implementation

How do all the pieces fit together? Does your frontend make requests to your backend? Where does your database fit in?

1. Client Interface Layer

• Technology: Flutter mobile application

• Functionality: Captures user input (e.g. chat, calendar, photo uploads) and communicates with backend via RESTful APIs secured using JWT tokens.

• Data: User queries, uploaded images, calendar metadata

2. API Gateway

• Technology: Node.js

• Functionality: Central entry point for the app; routes requests to appropriate microservices, handles auth tokens and request composition.

• Data: User preferences, login credentials, calendar events, and structured prompts

3. Microservices Layer

• Technology: Independent REST-based services

• Key Services:

• Recommendation Service: Provides personalised outfit suggestions using embedded style vectors and event context

• Calendar Service: Integrates calendar data for context-aware recommendations

• Chatbot Service: Interfaces with OpenAI API for natural language fashion advice

• Auth x User Profile Service: Manages authentication, user profiles, and state

• Data Flow: Internal data is exchanged as JSON over REST/gRPC

4. AI/ML Layer

• Technologies: Python (Hugging Face), TensorFlow, OpenAI GPT API

• Functions:

• Style Embedding Model: Extracts style vectors from user-uploaded photos

• Image Analysis Engine: Detects garment and attribute features

• OpenAI Chatbot: Composes natural language responses based on user input + system context

5. Cloud Infrastructure Layer

• Storage:

• AWS S3: Stores raw and processed images

• PostgreSQL (RDS): Stores user metadata, events, and logs

• Cloudinary: Delivers optimised image assets via CDN

• Hosting:

• Vercel/Render: Hosts frontend and backend microservices

• AWS RDS: Relational data persistence

System Flow Summary

Users interact with the mobile app to upload photos or submit chat queries. These requests are routed through the Node.js API gateway and handled by domain-specific microservices. AI/ML services process visual data and provide stylistic insights, while chat responses are dynamically generated via the OpenAI API. All data is persisted and optimized through cloud storage and CDN infrastructure.

https://drive.google.com/file/d/1hGVDYw3c6DJgFJp62M_LCLa-VVGuSW8E/view?usp=sharing

Challenges

What did you struggle with? How did you overcome it?

Idea generation. The project brief was very broad and so we had to think about a lot of different aspects such as our USP, technology stack and UX designs.

We overcame it by having several brainstorm sessions throughout the day which ensured that we all had a clear idea of our solution and how to implement it.

Accomplishments

What did you learn? What did you accomplish?

This Hackathon taught us that adaptability is just as important as technical skills.

We learned how to:

Collaborate quickly with new teammates.

Divide tasks efficiently with limited time and changing resources.

Pivot under pressure, stay focused, and still build something we’re proud of.

This experience highlighted the importance of communication, initiative, and being solution-oriented, especially in fast-paced environments like tech and startups.

Next Steps

What are the next steps for your project? How can you improve it

If we had more time, we would implement more features.

• Feature: “Style Twin” - AI connects users with similar fashion tastes for outfit inspiration and community styling.

• Swipe right if you like recommendation: AI learns and adjusts overtime

• Outfit analytics: outfit frequency, where was outfit worn

• Partnership with more brands to give users a wider selection of options

• Smart filters: search your closet by occasion, mood, last worn, weather suitability, etc

• Anonymous Fit Voting: Let user community vote between two outfits when you're unsure what to wear.

Team Roles and Contributions

We all acted as project managers and business analysts by asking several questions during the idea generation, ensuring our MVP was in scope and having open communication between the engineers and the business.

Chioma Uzor

Chioma Uzor worked on the front end development of the project using Figma.

Sarah Iruobe

Sarah Iruobe contributed to the market research and slide creation.

Bolaji Ogunnaike

Bolaji Ogunnaike was responsible for the User flow diagram and slide creation.

Sena Afi Vuvor

Sena Afi Vuvor focused on system architecture diagram and tech stack.

Fadekemi Adebayo

Fadekemi Adebayo worked on the front end development of the project using Figma.