Secure Terminal Background

Phaedra Tan //
Data Scientist | Cybersecurity Enthusiast

I'm a Data Scientist with a degree in Computer Science. I currently develop secure, full-stack web applications that deploy localized LLMs or government-approved AI instances for MOM. Outside of my primary work, I enjoy solving logic puzzles, from Sudoku to Python problems. Recently, I've gained an interest in using Wireshark to analyse passive traffic. I am looking to transition into cybersecurity roles where I can apply my hybrid background.

Find out more
Projects

Projects

Public // Decrypted

HR Alchemist: Job Fractionalization Tool

Launch System

A modular full-stack application that intelligently decomposes unstructured Job Descriptions into quantifiable, atomic Task Units for the Singapore Public Sector. The backend uses RAG pipe embeddings generated from PSD data to a secure, enterprise-grade OpenAI instance, ensuring the model is contextualised to Singapore Public Data. The frontend is created using React.js.

RAGVector DatabasePythonReact.js
Public // Decrypted

Kamui: Phishing Triage tool

Launch System

A specialized forensic tool to automate Level 1 analysis of suspicious email artifacts (.eml). The system recursively deconstructs nested email structures to isolate indicators of compromise, reconstruct delivery paths from raw headers, and generate dynamic threat scores. To ensure analyst safety, the tool utilizes memory-only processing and automated URL defanging to prevent accidental malware execution.

Phishing AnalysisEmail Header ForensicsThreat Intelligence & HashingSecurity AutomationDefanging
Classified // Encrypted

Predictive Modelling for Firm Insolvency

Access Denied // Classified

This project focuses on building a high-fidelity risk model using alternative data signals. I worked on engineering complex features from demographic and wage-trend data while integrating utility and payroll records to identify non-traditional distress indicators. Due to the nature of the project, it is difficult to share more information.

PythonRisk ModellingData EngineeringExplorative Data AnalysisData Visualization
Classified // Encrypted

Audit Report Classifier

Access Denied // Classified

An LLM-based tool that automates the analysis of sensitive audit reports. The model ingests reports and extracts and summarizes the key information, such as root causes and systemic issues. To adhere to strict data sovereignty requirements, the system was engineered for air-gapped deployment, utilizing fine-tuned open-source LLM models. The model achieved an average 90% accuracy rate in categorizing audit findings, and outputs the data into CSV and a visual dashboard.

LlamaIndexLLMs (Qwen, Phi)Air-Gapped DeploymentNLP
Certifications

[SECURE_STORAGE: VERIFIED_CREDENTIALS]

Certified in Cybersecurity (CC)

[CA]: ISC2
Access ControlSecurity OperationsIncident ResponseNetwork SecurityBusiness Continuity
[EFFECTIVE]: January 2026 - January 2029

Introduction to Vulnerability Management

[CA]: Security Blue Team
Vulnerability scanningAsset DiscoveryCVSS
[EFFECTIVE]: January 2026

Introduction to Digital Forensics

[CA]: Security Blue Team
[STATUS: UNVERIFIED]

Introduction to Network Analysis

[CA]: Security Blue Team
WiresharkPacket AnalysisInternet Protocol Suite (TCP/IP)Network Traffic AnalysisPCAP Investigation
[EFFECTIVE]: January 2026
Contact

Let's Connect

Since you stayed to the end, something about me must have stood out to you. Feel free to reach out via:

linkedin.com/phaedratan
phaedratan1@gmail.com
96375975
Return to Base