MyStudyPal — AI Study Buddy for Irish Students

MyStudyPal replaces generic, unstructured study methods with a targeted, curriculum-specific AI tool. It helps students actively identify weak points, generate concise revision notes, and test their knowledge against real Irish exam standards. The platform is backed by a continuously updated past-paper dataset, a high-performance Rust file server, and a fully containerized 7-service infrastructure.

Role: Deployment & Automation Engineer Scope: Full-Stack AI Educational Platform Focus: Microservices Deployment + Container Orchestration + Curriculum Data Automation

React HTML NestJS Rustfs PostgreSQL Redis Puppeteer Node.js Docker Dokploy

// overview

vision

Replace generic, unstructured study methods with a targeted, curriculum-specific AI tool that helps students actively identify weak points, generate concise revision notes, and test their knowledge against real exam standards.

audience

Students preparing for the Irish curriculum (Leaving Cert, Junior Cert) who need instant, accurate, exam-aligned feedback — as well as platform administrators managing content and user data.

outcome

Production-deployed platform serving students and admins efficiently. Highly available backend system driven by NestJS, Postgres, and Redis. Complete automation of past-paper data updates via Puppeteer. Clean, centralized infrastructure monitoring using Dokploy.

// problem

Students preparing for the Irish curriculum lacked a dedicated platform that provided instant, accurate feedback. Generic AI tools frequently hallucinate or fail to understand the highly specific marking schemes required for subjects like Irish, Science, and Maths. Core challenges included: (1) Curriculum Alignment — creating an AI that understands specific Irish exam grading criteria rather than just generic knowledge; (2) Media Processing — students need to submit screenshots of complex questions without lag, requiring an optimized file system; (3) Complex Deployments — managing 7 isolated components required a cohesive, easy-to-maintain infrastructure strategy; (4) Data Acquisition Pipeline — continually feeding the platform with the latest past papers manually was unscalable.

// solution

The backend is powered by a NestJS server acting as the core API gateway for user authentication, AI request routing, and data serving. Two React dashboards serve students and admins respectively, backed by a static HTML landing page optimized for conversion. PostgreSQL handles relational data (user progress, quiz results, notes), Redis provides high-speed caching and session management, and Rustfs serves as a dedicated high-performance file server for all media uploads. A custom Puppeteer automation script autonomously scrapes past paper data from official Irish curriculum sources and injects it directly into the NestJS backend, eliminating all manual data entry. Every application tier is Dockerized with production-optimized Dockerfiles and orchestrated through Dokploy for centralized lifecycle management.

// my_role

Deployment & Automation Engineer

› Designed and managed the entire microservices deployment strategy across 7 containerized services
› Containerized all 7 microservices using Docker with lightweight, secure, production-ready images
› Orchestrated the platform using Dokploy, configuring environment variables, internal networking, and volume persistence for Postgres, Redis, and Rustfs
› Engineered a Puppeteer web-scraping pipeline to extract past paper data autonomously and pipe it directly into the NestJS backend
› Ensured high availability, continuous uptime, and streamlined deployment workflows across all frontend and backend applications

// tech_stack

frontend

React, HTML

backend

NestJS

mediaServer

Rustfs (Rust)

databases

PostgreSQL, Redis

automation

Puppeteer, Node.js

infrastructure

Docker, Dokploy

// key_features

› AI Study Buddy with marking-scheme-aligned answers for Maths, Science, English, and Irish — supports direct image uploads
› Notes Generator that instantly produces clean, structured revision notes strictly based on the Irish curriculum
› Smart Quizzes dynamically generated by AI targeting weak areas, with instant correction and topic tagging
› Progress & Insights dashboard featuring an XP level system, weekly improvement statistics, and a strength/weakness heatmap
› Automated Puppeteer scraping pipeline for continuous past paper ingestion from official Irish curriculum sources
› 7 containerized microservices: Admin Dashboard, User Dashboard, Landing Page, NestJS Server, Rustfs, Postgres-DB, Redis-DB
› High-performance Rust-based file server (Rustfs) for fast, secure student question and image uploads
› Centralized infrastructure monitoring and lifecycle control via Dokploy dashboard

// contributions

› Designed and deployed full 7-service microservices architecture covering frontend, backend, databases, and file storage
› Wrote production Dockerfiles for every service tier, optimized for minimal image size and security
› Configured Dokploy orchestration: inter-service networking, environment variable management, volume mounts, and health checks
› Built Puppeteer automation script for autonomous past paper scraping, structuring, and backend injection
› Set up Rustfs as a dedicated Rust-based media server for low-latency student file and image uploads
› Maintained continuous uptime and deployment pipeline across all platform components

// challenges_&_solutions

Complex Multi-Service Deployments

Problem: The platform required 7 distinct services — including varied tech like React, NestJS, and Rust — to communicate securely in production without massive manual configuration overhead.

Solution: Utilized Docker to containerize each service independently, then orchestrated the ecosystem using Dokploy. This allowed for centralized logging, automated network routing, and seamless environment variable management across all tiers.

Scaling AI Context Data

Problem: The AI Study Buddy required continuous feeding of highly specific Irish past paper data to provide accurate, localized answers. Manual updates were unscalable and error-prone.

Solution: Built a robust automation script using Puppeteer that scrapes official curriculum data, structures it, and automatically injects it into the NestJS backend — entirely eliminating manual data entry and keeping AI responses current.

// impact

7 Containerized Microservices

Robust, isolated, and independently deployable services covering every platform concern — frontend dashboards, backend API, databases, media server, and landing page — all orchestrated through a single Dokploy interface.

100% Automated Data Ingestion

Past paper data is scraped, structured, and injected into the backend autonomously via Puppeteer — ensuring AI responses always reflect the latest official Irish curriculum content without any manual intervention.

Unified Deployment Control

Centralized infrastructure management via Dokploy dashboard provides clear visual insight and lifecycle control over all running containers, environment configs, and service health.

High-Accuracy AI Responses

Strict marking-scheme alignment — enforced through curriculum-specific data feeding — ensures the AI Study Buddy delivers accurate, exam-relevant answers rather than generic responses.

// what_i_learned

This project gave extensive hands-on experience in managing complex deployment architectures and container orchestration. Using Dokploy alongside Docker simplified what would have been a tangled web of microservices into a clean, highly maintainable infrastructure. Building the Puppeteer pipeline highlighted the critical intersection between DevOps, automation, and AI accuracy — ensuring the backend always had the most accurate, curriculum-specific data without manual intervention proved to be just as vital as the AI models themselves.