Project Story: SchoolDaddy – AI Agent for Graduate Student Success

Inspiration

As a graduate student juggling coursework, research, and personal commitments, I often struggled to keep track of my own performance and well-being. I realized that many students face similar challenges—limited time, multiple deadlines, and difficulty identifying early signs of academic or personal stress.

This inspired me to create SchoolDaddy, an AI-powered platform that acts as a personal academic and well-being assistant. The goal is to help students proactively monitor their performance, prioritize tasks, and receive actionable insights before challenges escalate.

What it does

SchoolDaddy is a unified platform that:

Monitors student performance and well-being over time.
Identifies at-risk students using predictive models.
Provides actionable insights and personalized recommendations.
Supports natural language queries so students can ask, for example:
> “Which tasks should I prioritize this week to avoid falling behind?”
Visualizes progress trends and risk factors through interactive dashboards.

In essence, it helps students like me balance multiple commitments and make data-driven decisions about time and effort allocation.

How we built it

The project leverages a multi-layer AI and data ecosystem:

Synthetic Data Generation – Simulated academic schedules, task deadlines, well-being scores, and engagement metrics using Python (pandas, NumPy, datetime, random).
Predictive Intelligence – RandomForestClassifier models track risk levels and predict performance dips based on behavioral and temporal trends. Integrated with MLflow for experiment tracking.
Causal Analysis – Planned integration with DoWhy/EconML to understand why a student might be struggling, rather than just flagging them.
Databricks Integration – Centralized all data and ML pipelines in Databricks Lakehouse. Delta tables and PySpark handle large-scale processing efficiently.
AI Agent & RAG Workflow – The AI agent allows natural language queries. Retrieval-Augmented Generation (RAG) fetches relevant context from the dataset and provides actionable insights.
Interface Layer – Gradio dashboards with Plotly visualizations give real-time feedback and task prioritization guidance.

This setup allows even a busy student to query their academic and well-being metrics without diving into raw data.

Challenges we ran into

Data Realism – Modeling realistic student behavior and stress patterns for synthetic datasets was tricky.
Feature Engineering – Identifying meaningful signals from engagement, assignment completion, and well-being metrics required multiple iterations.
System Integration – Combining Databricks pipelines, ML models, and the Gradio interface while ensuring real-time responsiveness was challenging.
Causal Inference – Understanding root causes behind performance dips is complex, especially when multiple factors (time constraints, personal issues, workload) interact.

Accomplishments that we're proud of

Successfully simulated a graduate student ecosystem with multi-tasking and time constraints.
Built predictive models capable of identifying students at risk of falling behind.
Developed an AI agent with natural language querying, making insights accessible to non-technical users.
Designed an interactive, cloud-ready dashboard that visualizes both academic performance and well-being trends.

What we learned

Managing and analyzing multi-dimensional student data provides actionable insights beyond raw grades.
AI and causal analysis can be applied to personal time and task management, not just institutional risk monitoring.
Integrating RAG workflows with Databricks enhances the context-awareness of AI responses for real-world decision-making.
Building a student-centric AI platform requires balancing prediction, explanation, and usability.

What's next for SchoolDaddy

Connect to live graduate student schedules and task-tracking apps for real-time recommendations.
Incorporate personalized time management guidance to help students prioritize based on deadlines and energy levels.
Add real-time intervention suggestions (e.g., study reminders, wellness check-ins) using the AI agent.
Enhance causal analysis to suggest actions that directly mitigate academic or well-being risks.
Expand the platform to support collaborative workflows for students working on group projects or research teams.