Overview

This project involved designing and implementing an end-to-end recruitment system combining candidate intake, automated CV screening, and centralized data management.

The system integrates Google Forms (frontend and database) with Google Apps Script (automation engine) to reduce manual workload and improve screening efficiency.

Problem

Recruitment processes relied heavily on manual CV screening, spreadsheet tracking, and inconsistent evaluation criteria. This resulted in slow turnaround times, high administrative workload, and limited visibility into candidate quality.

System Design

• Frontend: Public candidate submission form built using Google Form
• Data Layer: Centralized candidate database with structured fields and scoring logic
• Backend: Google Apps Script linked to Gemini API for CV parsing, data extraction, and automated scoring

How It Works

1. Candidate submits application through AppSheet form
2. Data is stored in a structured database
3. Google Apps Script extracts key information from CVs
4. System applies scoring logic based on experience, education, and skills against the applied Job description
5. Candidates are automatically ranked and categorized

System Architecture

Key Features

• Automated CV parsing and data extraction
• Standardized candidate evaluation framework
• Dynamic scoring and ranking system
• Centralized tracking across recruitment stages

Results

• Reduced manual CV screening effort by up to 90%
• Improved consistency in candidate evaluation
• Enabled faster identification of high-quality candidates
• Freed up recruiter time for interviews and decision-making

Documentation & Technical Details

• View Full System Documentation →

Lessons Learned

• Automation requires sustainability: The system relied on Gemini API tokens for CV screening. The project was discontinued after 3 months due to token limitations and budget constraints, highlighting the importance of aligning technical solutions with long-term operational support.
• AI improves speed, but not accuracy alone: While automation significantly reduced manual effort, human validation remained necessary. AI-generated results were cross-checked to ensure reliability.
• Fail-safe mechanisms are critical: Implemented random sampling on rejected candidates to ensure qualified profiles were not overlooked due to AI misclassification.
• Change management matters: Adoption of automation tools depends not only on technical effectiveness, but also stakeholder understanding and buy-in.

Back

← Back to Portfolio