Skill Forge
Mumbai University CE/IT ed-tech platform with a PYQ engine and Placement Readiness Score.
screenshot pending
Supply real screenshots before Phase 8
The Problem
Mumbai University Computer Engineering and IT students struggle to find structured study resources, syllabus-aligned Previous Year Questions (PYQs), and clear guidance on their placement readiness, leading to fragmented exam preparation and sub-optimal placement results.
The Solution
Developed a unified ed-tech platform featuring an automated OCR-driven PYQ indexer, semester-wise curated study resources, and an interactive placement readiness scoring dashboard that weights core computer science topics.
Key Features
- OCR PYQ EngineCustom parser to transcribe low-contrast university question papers into searchable databases.
- Placement Readiness ScoringMetric algorithms mapping academic scores and project completions to target tier readiness.
- Semester NavigationOptimized study folder index mapping resources to specific MU board syllabi.
Technical Architecture
- REST API LayerNode.js and Express server handling question fetches and dashboard metrics logic.
- MongoDB AggregationsDynamically weights test responses and calculates cohort statistics using indexing.
Key Engineering Challenges
Scanned question paper PDFs provided by MU are often noisy, skewed, or blurred. Solved by implementing a pre-processing pipeline that adjusted image thresholding, scaled contrast using Canvas APIs, and ran clean OCR segmentation.
Calculating placement scores for hundreds of concurrent students doing mock tests thrashed database CPU performance. Solved by utilizing cached score aggregation weights and running updates in micro-batches rather than on every telemetry event.
Key Metrics
Lessons Learned
Ed-tech platforms built for student environments must optimize data transfer sizes. Pre-splitting PDFs and caching question indices locally yields superior accessibility.