Rafii Ahmad Fahreza
Back to portfolio
AI / EdTech· 2026

Personal Learning Manager (PLM)

Agentic AI-Powered Learning System for self-regulated, adaptive education.

Next.jsSupabasepgvectorGemini APIn8nGoogle Calendar API
Personal Learning Manager (PLM)

Overview

Traditional study tools treat a 50-page reading and a 10-minute quiz the same way, offering no real help with the actual bottleneck of self-regulated learning: organizing, digesting, and scheduling material.

The Personal Learning Manager (PLM) closes this gap by transforming static study materials into dynamic, personalized learning experiences. By combining Agentic AI workflows with Retrieval-Augmented Generation (RAG), the system acts as a persistent learning companion.

Designed specifically for students managing heavy academic loads, PLM bridges the gap between active study time and everyday constraints, making personalized learning practical.

Key Architecture & Features

PLM integrates powerful AI agents that execute complex multi-step workflows. Uploaded documents are parsed, indexed in a PostgreSQL database with pgvector, and queried dynamically using RAG techniques to construct a custom knowledge base.

An automation layer powered by n8n orchestrates background workflows, coordinating between embedding models, large language models (LLMs), and external services like the Google Calendar API for adaptive scheduling.

Learning Space & RAG Assistant

Organize documents into dedicated subject spaces. Ask questions, draft summaries, and retrieve information with a context-aware AI tutor.

Adaptive Learning Paths

AI analyzes course syllabi to generate step-by-step learning paths that adjust automatically based on quiz performance.

Active Recall & Quizzing

Automatically generates personalized, concept-based quizzes to test retention, reinforcing weak areas on the fly.

Adaptive Scheduling & Calendar Sync

Calculates required study hours and syncs events with Google Calendar, shifting plans automatically if real-world schedules change.

Project Status & Target Outcomes

Active Dev

Status

Gemini/OpenAI

AI Backend

n8n Agent

Orchestration

Currently under active development, this system is being built to serve as a research prototype exploring self-regulated learning frameworks powered by AI.

The goal is to test the efficacy of Agentic AI in helping students reduce study-planning overhead, increase retention through active recall, and maintain consistent study habits.