AI Powered Platform for Developing Students' Potential

Project Summary: Software platform that helps 40,000 students in the USA to discover and unleash their potential. The platform is using artificial intelligence to help students explore their potential and guide them in developing it. I was the technical lead of the data science team.

Project Duration: 9 months

Team Size: 30 team members, 5 data scientists

Industry: Education

Customer

Independent school district (ISD) in the USA with about 40,000 students.

Business Case

The ISD wanted to help K-12 students discover and unleash their full potential using the power of data and artificial intelligence. The district uses multiple education platforms which collect a lot of data about each student. This data makes it possible to extract valuable insights and train machine learning models. The challenge is that the data is unstructured (mostly text) and scattered across different locations.

Solution

Our team built a software platform in which students can explore and learn more about their interests, natural abilities, and potential in different areas of life. Multiple artificial intelligence algorithms help students in their exploration by automatically identifying and suggesting areas for exploration and guiding them in their potential development.

The exceptional user experience of the platform keeps students engaged. The high engagement provides a stable feedback loop for the machine learning models which improve over time to give ever more relevant suggestions and guidance to the students.

The machine learning algorithms used in this project are very diverse. Some of the used techniques are:

  • Edge prediction in graph data structures
  • Matrix factorization for collaborative filtering
  • Pre-trained HuggingFace transformer models for natural language processing

Tools & Technologies

  • Google Cloud Platform
  • Google Cloud BigQuery
  • Google Cloud Dataflow
  • Google Cloud Composer
  • Google Data Studio
  • Python
  • NetworkX
  • HuggingFace

My Role & Responsibilities

I was technical lead of the data science team working on the machine learning part of the project. My responsibilities were:

  • Technical consultancy
  • Facilitation of the cross-team communication
  • Support for the product management team in building the product roadmap
  • Defining technical requirements and work estimations
  • Data science team management
  • Data analysis
  • Development and deployment of data processing pipelines
  • Machine learning research & development
  • Productionalization of the machine learning solution

Results

The platform was launched in production and is currently used by 40,000 students in the USA. The platform is helping students to discover and unleash their potential. The platform is also helping teachers to better understand their students and provide them with more personalized guidance.