Capstone Project Team

Meet the Developers.

We are the co-developers behind AI-Grader. Together, we designed and engineered the full-stack web dashboard, the Android mobile application, and the deterministic grading engine to make academic integrity auditable, transparent, and fair.

Zyril B. Lumbay

Zyril B. Lumbay

Co-Developer & Full-Stack Web Engineer

Engineered the complete web application from the ground up — owning every layer from database schema to production UI. Built the frontend dashboard with a component-driven, responsive design system, developed secure PHP backend APIs for authentication and role-based access control, and designed the MySQL architecture with normalized schemas and indexed queries. Implemented the core deterministic math engines — TF-IDF, Cosine Similarity, and Flesch-Kincaid — powering keyword extraction, plagiarism scoring, and readability analysis. Integrated Groq and SerpApi REST APIs for AI-driven essay evaluation and web-source cross-referencing, with server-side key protection and resilient error handling. Hardened the stack with CSRF tokens, session authentication, and login-attempt throttling, delivering a unified grading pipeline that fuses algorithmic scoring, AI feedback, and plagiarism detection.

HTML/CSS JavaScript PHP AUTH MySQL REST APIs System Architecture
Adry P. Tagamolila

Adry P. Tagamolila

Co-Developer & Mobile Engineer

Spearheaded the cross-platform integration and end-to-end mobile development. Built the Android application pipeline from the ground up in Android Studio, including project structure, build configuration, and deployment workflow. Implemented the ML Kit OCR scanning features for digitizing essay submissions via device camera, handling image preprocessing, text recognition, and structured data extraction. Developed and secured the REST API integration layer, enabling real-time data synchronization between the mobile app and the web database through authentication, request validation, and resilient network handling. Collaborated with the web team to define API contracts and standardize data schemas, ensuring consistent behavior across platforms.

Java Android Studio OCR / ML Kit API Integration
Project Scope

The 6 Pillars of AI-Grader

From deterministic math to mobile OCR, here is the complete suite of features we engineered for the institution.

01 / Plagiarism

travel_explore Peer Plagiarism Scan

TF-IDF and Cosine Similarity to detect internal student collusion and highlight copied sentences.

02 / Internet

public Internet Source Scan

Google SerpAPI integration to check essays against live web sources for exact matches.

03 / Authorship

verified_user WriteProof™ Verification

Tracks keystrokes, paste events, and tab switches to verify the student typed their own essay.

04 / Quality

edit_note Mechanics Review

Evaluates spelling, vocabulary richness, sentence variety, and readability.

05 / Relevance

target Topic Alignment

Groq AI comparison of the essay against the teacher's uploaded lesson content.

06 / Physical

document_scanner Paper Essay Scanning

Android-based OCR that converts handwritten exams into digital text for analysis.

Ramon Magsaysay Memorial Colleges - Marbel, Inc.
College of Information Technology Education