Enlear HRMS — Serverless Enterprise Platform

Overview

Enterprise HRMS is a cloud-native Human Resource Management System designed to streamline employee management, attendance, leave, and device allocation. It provides role-based experiences for Administrators, Managers, and Employees while integrating workplace services such as Google Calendar, Google Chat, Slack, and GitHub. Built with a serverless AWS architecture using AppSync, Lambda, DynamoDB, SQS, and Cognito, the system is designed for secure and scalable operations. An Amazon Bedrock-powered AI chatbot further enhances the platform by providing intelligent assistance for HR-related tasks.

The Problem

Growing organizations face significant operational overhead when scaling HR processes:
1.Challenges due to fragmented HR and operational systems
2.Attendance, leave management, and productivity tracking operated independently
3.Manual attendance logging caused inaccuracies and delays
4.Leave approvals were time-consuming and lacked real-time visibility
5.No centralized system to track company devices and reported issues
6.Difficulty correlating developer work output (GitHub activity) with attendance records
7.Limited analytical insights for managers and administrators

Goals

1.To design and develop a custom Human Resource Management System for Enlear
2.Automate HR workflows and reduce administrative overhead
3.Improve transparency, efficiency, and decision-making
4.Provide a scalable and secure platform tailored to Enlear’s workflow

My Role & Key Decisions

1.Device management module for all 3 roles( including device assignment, device issue managment)
2.Integrated AWS subscription (graphQL) for internal system notifications with using AWS SQS.
3.Implemented browser session-based chat history for chatbot to retain conversations until the employee closes the tab or logs out.
4.Create test cases using vitest.

System Architecture

The architecture separates client interaction from cloud execution using a serverless event-driven model:

What I Learned

• Designing serverless architectures demands thinking in asynchronous boundaries: delegating side-effects to queues like Amazon SQS keeps client interactions snappy and resilient.
• Strong GraphQL typing paired with Vitest unit tests provides unmatched confidence when modifying complex business calculation engines.

Future Improvements

• Implement biometric mobile check-in with geofencing validation.
• Add advanced predictive attrition modeling based on historical leave patterns and feedback cycles.