01 In Progress

Network Attack Visualiser

Defensive cybersecurity platform that detects TCP port scans and SYN-flood attacks in real time inside an isolated, authorised Docker lab, and explains every alert with transparent, confidence-scored heuristics.

  • Three ingestion modes: synthetic events, PCAP replay, and authenticated live capture
  • Two heuristic detectors with clock-injected, deterministically testable logic
  • Metadata-only capture, no packet payloads ever stored, by schema design
3Ingestion modes
0Payload bytes stored
GPL-2.0Open source
Python FastAPI Scapy React Docker
The Network Attack Visualiser dashboard: live alert feed showing a synthetic SYN-flood and port-scan detection, protocol distribution, traffic timeline, and the alert investigation pane
02 Shipped

Caissa Chess Overlay

Real-time desktop computer-vision application that detects chessboards from screen pixels, reconstructs legal positions, and streams local Stockfish analysis through a responsive PyQt6 overlay.

  • Locates and reads any board straight from screen pixels, no site API
  • Local Stockfish analysis streamed live, first read in ~100 ms
  • Self-contained Windows build, fully offline, nothing sent anywhere
~100 msFirst move
100%Offline and local
GPLv3Open source
Python OpenCV PyQt6 Computer Vision Stockfish
Caissa floating beside a live chess game, showing the evaluation and the best move F3 to G5
03 Live

OMI Multiplayer

Production-deployed real-time multiplayer game showcasing WebSockets, secure server-authoritative architecture, Cloudflare DNS/TLS, and Northflank deployment.

  • Real-time Socket.IO multiplayer with a server-authoritative state machine
  • Host and Origin validation, rate limiting, CSP hardening
  • Continuous deployment via GitHub, Northflank, and Cloudflare
LiveIn production
90sReconnect grace
20/sRate-limited events
Node.js Socket.IO Cybersecurity Networking Northflank
The OMI join screen: name entry, Join Game and Leaderboard buttons, on a felt-green table background
01/03 Selected work