Projects
Security work & home labs
Hands-on builds across blue-team monitoring, network security, applied cryptography, and access control. Where a project lives on GitHub, the repo is linked.
Blue Team · Detection Engineering
Active Directory Security Monitoring Lab
Built a virtualized Windows Server 2019 environment running Active Directory Domain Services, DNS, and file sharing for domain management. Applied Group Policy audit settings to track logon events and privilege use across domain-joined Windows clients — the foundation for detecting lateral movement and privilege escalation in an enterprise domain.
Home lab · Jun 2025
Network Security · DNS
Pi-hole DNS Sinkhole & Network Filtering
Deployed Pi-hole on a Linux-based home network to block ad, tracker, and known-malicious traffic at the DNS layer across every connected device. Used the built-in admin portal to monitor query volume, inspect resolution behavior, and manage blocklists from a local IP portal — a practical look at DNS as both a control point and a telemetry source.
Home lab · May 2025
Applied Cryptography · Privacy
ShadowMsg
An encrypted messaging application built for privacy. It operates without accounts or server-side logging, and sessions and messages terminate automatically — minimizing the data that exists to be compromised in the first place. An exercise in threat-modeling a communication tool and designing for data minimization.
Access Control · Multi-Tenancy
LockedInJobs
A collaborative job-application tracker with strict per-account data isolation. Each user's applications, status history, and aggregate statistics stay scoped to their own account — a real-world exercise in authentication, authorization, and preventing cross-tenant data leakage.
Threat Intelligence · Built with Claude
Threat-Intelligence Dashboard
A Next.js dashboard that aggregates cybersecurity advisory and news feeds — The Hacker News, Krebs on Security — into a single monitored view alongside schedule and task data. API keys live in environment variables and are never committed; deployed on Vercel. A lightweight personal threat-intel and OSINT feed.
AI Security · Built with Claude
Autonomous Agent with Least-Privilege Guardrails
An unattended AI agent constrained by a hard security boundary: a single authorized account, an explicit allow/deny scope, mandatory verification before any state-changing action, and an immutable run log for every execution. A study in applying least privilege, segregation of duties, and audit trails to autonomous systems.
Red Team / OSINT Research · Built with Claude
Red Team & OSINT Knowledge Base
A self-built 128-note reference hub covering MITRE ATT&CK, command-and-control infrastructure, Active Directory attack and defense, and OSINT investigation methodology. Every offensive technique is paired with a “Detection & Defense” section and ATT&CK technique IDs. Scoped strictly for authorized lab and CTF use.
Web Tool · JavaScript
Environmental Economics Tool
A browser-based calculator for environmental economics: equilibrium problems, tax calculations, and present-value analysis with automated visualizations. Runs entirely client-side — no data leaves the browser.
Machine Learning · Python
2048ai
A self-taught AI that plays 2048, trained with an n-tuple value network via TD(0) reinforcement learning and expectimax search for move selection. A Tkinter desktop app handles training, live progress charts, and model management, and can read the board off a live browser game at 2048.org to play it directly. Reaches the 2048 tile in 96% of games and 8192 in over half.