<< BACK TO PORTFOLIO
CONFIDENTIAL // INTERNAL USE ONLY

Engineering Documentation

Project Architecture & Technical Implementation Details

MODULE 1: NET_PROBE

Production
💡 Core Concept
A tool to measure connection quality between a user's location and AWS Singapore Region. It helps identify if high latency is caused by the user's ISP routing or the AWS network itself.
Net Probe Architecture Diagram
Figure 1.1: Event-driven Serverless Architecture (Region: ap-southeast-1)

Tech Stack

  • Ingestion: AWS API Gateway
  • Compute: AWS Lambda (Python)
  • Database: DynamoDB (Logs storage)
  • Security: IAM Roles (Least Privilege)

How it works

STEP 1 User sends a request from the browser.
STEP 2 Lambda triggers a network probe (Ping/TCP) to external targets.
STEP 3 Results (Latency/Location) are saved to DynamoDB.
STEP 4 Analysis is sent back to the user instantly.

MODULE 2: FACE_BIOMETRIC

Refactoring
💡 Core Concept
An experiment combining Face Matching (Rekognition) with AI Text Generation (Bedrock). Instead of just saying "Match/No Match", the system uses AI to generate a readable report explaining the similarity confidence.
Biometric Architecture Diagram
Figure 2.1: AI Pipeline integrating Rekognition & Bedrock

Tech Stack

  • Orchestration: AWS Lambda
  • Vision AI: Amazon Rekognition
  • Gen AI: Amazon Bedrock (Claude 3 Haiku)

How it works

STEP 1 Image is uploaded via API Gateway.
STEP 2 Lambda sends image to Rekognition for vector analysis.
STEP 3 Bedrock generates a summary report based on the data.

MODULE 3: SOS_BEACON

Mission Critical
💡 Core Concept
A personal emergency alert system. When triggered, it captures precise GPS coordinates and sends them to a pre-configured list of family members/admins.

Note: I use a verified list of recipients (via AWS SNS Sandbox) to ensure 100% delivery reliability and prevent messages from being flagged as spam by email providers.
SOS Architecture Diagram
Figure 3.1: High-availability Broadcast System

Tech Stack

  • Telemetry: HTML5 Geolocation API
  • Broadcast: Amazon SNS (Fan-out)
  • Logic: Lambda (Deep-linking generation)

How it works

STEP 1 Browser captures GPS Coordinates.
STEP 2 Signal is sent securely to AWS.
STEP 3 SNS broadcasts SMS/Email to the verified list instantly.

SECURITY NOTE

To follow AWS Security Best Practices (preventing credential leaks), the GitHub repository is currently set to PRIVATE.

I am happy to show the code and explain the implementation details in person during the interview.

END OF DOCUMENT // REGION: AP-SOUTHEAST-1