Projects

I like to build things. Here’s some of them:

Notes visualiser #

I’ve been writing short notes down about thoughts, ideas, things to do etc. for a few years and so now I have over 2000. I wanted a way to visualise these and see how notes are related to other notes and the clusters of where a lot of my notes are. For this, I first processed and cleaned the data, then I used the BERT Sentence transformer to convert each note into a vector. I then used UMAP to dimensionally reduce this vector into six dimensions, 3 dimensions represented the note in space and 3 dimensions represented the colour of the note. I used the 3d-force-graph libarary to visualise it. Here’s the final result:

Faast-go #

Adopting the philosophy of Infrastructure as Code, an open source simple enumeration tool for recon and payload sending through YAML rather than the command line. Using worker pools for parallel computing and written in Golang. Github repo

Guided AI #

Initially born from a hackathon project me and my friend worked on. We ended up winning some prize money and so thought to continue the idea. Closely integrates biometric data while doing physical activities such as running, pushups or meditation to give personalised feedback as if you had a personal coach with you. It uses text-to-speech technology (initially using Elevenlabs but now fine tuned our own model) to give realistic and motivational voices such as David Goggins so it sounds like he’s with you while you are training. Watch our video submission here.

Touch grass #

This is the project I’ve been working on the longest. Deals with all layers of the stack and technology. Hardware configuring the Raspberry pi, setting up scripts and services to run on the Pi, implementing the Go client that runs on the Pi, ansible scripting for setup, custom protocol on top of TCP/TLS, backend server with authentication + routing + communication for HTTP and custom protocol, frontend with NextJS, real time notifications with websockets and pg_notify, infrastructure set up and configured with Docker and docker compose. More details in blog post

graph TB
    subgraph User["User"]
        HD[Hardware Device
Raspberry Pi]
        Phone[Smartphone]
    end

    subgraph Server["Server Infrastructure"]
        NGINX[NGINX Reverse Proxy]
        NextJS[Next.js Frontend]
        GO[Go Application]
        DB[(PostgreSQL
with PostGIS)]
    end

    subgraph External["External"]
        WIFI{WiFi Hotspots}
    end

    %% Connections
    HD --"Connects to"--> WIFI
    HD --"Sends data
(TCP/8888)"--> GO
    Phone --"Accesses website
(HTTPS)"--> NGINX
    NGINX --"Proxies requests"--> NextJS
    NGINX --"Proxies API requests"--> GO
    NextJS --"Fetches/Sends data"--> GO
    GO --"Stores/Retrieves data"--> DB

    %% Hardware Device Details
    subgraph Hardware["Hardware Device Details"]
        BASH[Bash Script]
        GOCLIENT[Go Client]
    end

    HD --- Hardware
    BASH --"Runs"--> GOCLIENT

    %% Styling
    classDef server fill:#4a5568,stroke:#e2e8f0,stroke-width:2px,color:#e2e8f0;
    classDef external fill:#2c5282,stroke:#e2e8f0,stroke-width:2px,color:#e2e8f0;
    classDef hardware fill:#2f855a,stroke:#e2e8f0,stroke-width:2px,color:#e2e8f0;
    classDef phone fill:#b7791f,stroke:#e2e8f0,stroke-width:2px,color:#e2e8f0;
    classDef default color:#e2e8f0;

    class NGINX,NextJS,GO,DB server;
    class WIFI external;
    class HD,BASH,GOCLIENT hardware;
    class Phone phone;

    %% Set background color
    %%{init: {'theme': 'dark', 'themeVariables': { 'darkMode': true, 'background': '#1a202c', 'primaryColor': '#e2e8f0', 'secondaryColor': '#718096', 'tertiaryColor': '#4a5568' }}}%%
  

Seasame fox 🦊 #

A potential way for mass surveillance on a budget. Unfortunately iPhone and android have MAC address randomisation and rotation so no longer possible :( Sorry NSA

Arduino controlled LED lights #

Integrated with my room for controlled and programmable LED lights. Reverse engineered my TV remote’s HEX codes to be able to control the LEDs via the remote. Github repo

Homemade breathalyser #

Fun to bring to parties to see how drunk you are, using Arduino and 3d printed chassis and custom LED screen. Github repo

Simulating gambling tactics #

Written in python and then optimised in C++ for performance. Used to test my hypothesis about the Martingale strategy in Roulette and the mathematics behind it. More details in blog post

Bookmark Manager #

This was my first big project when I was around 17. I had no idea what I was doing. Everything was plain javascript and css (and jquery). It was for my personal use case and then I shared it as an extension on the chrome web store. It has 250+ daily active users, 2.5k total installs, 270k total impressions. Github repo