DevBucket

Inspiration

As a community-driven individual, I understand the challenges faced by students in my community, especially those unsure about their career paths. Many students spend unproductive time in college, while those with passion lack a structured roadmap for skill development to become job-ready. Furthermore, in our tier-3 colleges, internship opportunities are scarce due to remote locations. To address these challenges, my friend and I built DevBucket, a solution to help students become job-ready.

What it Does

DevBucket is a machine learning-based web application that empowers students to build their core skills and become job-ready. The platform offers:

Mentorship: Students receive guidance from a machine learning model and professional community members.
Roadmaps: A structured learning path across different domains.
Internship Opportunities: Real-world projects to enhance skills.

Our platform helps students access the latest learning resources and hands-on projects to refine their technical abilities.

How We Built It

DevBucket utilizes a microservice architecture. I focused on Python, Flask, and Azure Machine Learning, while my friend handled the front-end with Java. Here's a breakdown of our architecture:

ML/AI: Using Azure ML and Azure AI services, I deployed a Random Forest machine learning model, creating a REST endpoint. This endpoint provides career mentorship guidance based on user data. For the chatbot, we used GPT-3.5-Turbo along with RAG and OpenAI embeddings to retrieve information from books uploaded by students.
Web Application: The backend was developed using Flask and Springboot. The roadmap and internship sections were built in Java, while the NoCodePortfolio section was handled in Python (Flask). Data is stored in MySQL, allowing dynamic fetching of roadmap details.

Challenges We Faced

API limitations: The OpenAI key had restrictions on queries per minute and token limits for text generation.
Azure ML Integration: Deploying models on Azure ML and integrating them with Java applications was challenging.
Chatbot Integration: Linking the chatbot seamlessly to the application posed technical hurdles.
Dynamic Data Fetching: Retrieving data dynamically from the MySQL database for the roadmap section was complex.

Accomplishments

We are proud of creating DevBucket, a complete solution for students at smaller colleges. Our key accomplishments include:

Comprehensive Solution: Providing mentorship, career paths, and internship opportunities.
Smart Chatbot: Quickly retrieves information from uploaded books.
Team Collaboration: Effective teamwork and task management.
High Accuracy: Azure ML services provide career guidance with 96% accuracy.

What We Learned

This project offered invaluable learning experiences, including:

Azure ML Expertise: Gained proficiency in deploying models and creating endpoints.
AI Integration: Successfully implemented GPT-3.5-Turbo, RAG, and OpenAI embeddings.
Team Collaboration: Worked professionally as a team, solving challenges and integrating components in real-time.

What's Next

To scale and improve DevBucket, we plan to:

Enhance Accuracy: Collect more data to improve our machine-learning model across various domains.
Build Partnerships: Collaborate with organizations to provide more internship opportunities.
Expand Reach: Leverage social media for marketing and explore funding opportunities to scale our project.
Ensure Security: Implement robust security measures to safeguard user data.

Technology & Languages

[ ] JavaScript
[X] Java
[ ] .NET
[X] Python
[X] AI Studio
[ ] AI Search
[ ] PostgreSQL
[ ] Cosmos DB
[ ] Azure SQL

Project Repository URL

https://github.com/savageheart/AI-Classroom

Deployed Endpoint URL

No response

Project Video

https://drive.google.com/file/d/1At9fZYfP25Zo6IZNulAB4EPrwBNnetXh/view?usp=sharing

Team Members

savageheart, habibrahman29

microsoft / RAG_Hack

Empowering Students with DevBucket: A Comprehensive Path to Career Readiness #136

Project Name

Description