Apical

Problem Space

The platform needed to support both structured English test preparation and dependable online evaluation workflows. That meant joining content delivery, test execution, scoring support, and frontend responsiveness in one cohesive product.

What Was Built

• Designed and developed a comprehensive Pearson Test of English preparation and online training platform.
• Built RESTful Laravel APIs and ensured smooth integration with a Nuxt-based frontend experience.
• Integrated Azure Cognitive Services to automate and evaluate test scores with high accuracy.

Engineering Approach

Laravel handled the application and API layer, while Vue.js and Nuxt supported the learner-facing interface and delivery flow. MySQL schema and query design were optimized to keep performance fast and scalable as users moved through training and test-related workflows.

Test Workflow Diagram

• Step 01: Candidate registration, package selection, and session scheduling were handled through the Nuxt frontend with validation against backend availability rules.
• Step 02: Practice modules, mock tests, and timed exam sessions were delivered through dedicated APIs so reading, listening, speaking, and writing flows stayed isolated and trackable.
• Step 03: User responses were persisted to Laravel services, normalized by test type, and queued for score preparation and administrative review.
• Step 04: Final score data, attempt history, and feedback summaries were returned to the learner dashboard and training staff portal in one consistent reporting flow.

Scoring Logic

The scoring layer combined rule-based validation with Azure Cognitive Services support for speech- and language-oriented evaluation tasks. Objective sections such as multiple-choice or structured answers could be scored immediately, while speaking and writing submissions followed a controlled pipeline where AI-assisted scoring outputs were mapped into the application’s internal score model before being surfaced to learners or instructors.

• Separated raw answer capture from calculated score records to keep grading reproducible.
• Stored per-section scoring metadata so staff could trace how a final result was assembled.
• Allowed administrative review workflows for edge cases before publishing sensitive result summaries.

API Architecture Notes

The Laravel API was organized around clear domain boundaries: authentication and user profiles, course and content delivery, exam session management, answer submission, scoring, and reporting. This kept the Nuxt frontend lightweight while allowing each exam-related workflow to evolve without tightly coupling every screen to one large backend controller layer.

• Authentication endpoints handled learner and staff access with role-aware responses.
• Exam APIs managed session state, timers, question payloads, and answer persistence.
• Scoring and reporting endpoints exposed summarized results for dashboards, review queues, and training analytics.