Notes

A review of key developments in the synthetic data landscape over the past few years, driven by advances in generative AI and falling costs, and a practitioner's perspective on the opportunities and challenges ahead.

A guide on how to execute ComfyUI workflows as standalone scripts.

Walkthrough on streaming structured objects to create progressively updating interfaces with FastAPI and Next.js.

A survey on the different methodologies used to generate structured output from LLMs, from model fine-tuning, to domain specific language, and schema engineering.

A concise guide on building SQLite WASM on Ubuntu Linux with custom extensions. Run SQLite in the browser and enable new possibilities by providing an interface to other C libraries through custom extensions.

A review of the things I have learned from building in the open over the past year. Thoughts and reflections on what it takes to grow a project and the difficulty translating open-source success to commerical success.

An exploration of markdown and HTML syntax trees. Documenting my experience creating rehype-prism-plus, a syntax highlighting plugin that creates pretty code blocks.

A good project is only one part of the puzzle. Getting stars is really all about marketing and promoting it. A guide on growth hacking a Github project.

Learn Julia by implementing Schelling's famous segregation model. You will see many similarities to Python - no types need to be specified (it's a dynamic language) and pick up some nice syntactical properties of Julia.

A revised benchmark of graphs / network computation packages featuring an updated methodology and more comprehensive testing. Find out how Networkx, igraph, graph-tool, Networkit, SNAP and lightgraphs perform

How we engineered a large scale label propagation algorithm at Cylynx

The serverless way - using Google Cloud Platform to deploy simple machine learning models via Cloud Run. A fun weekend project that analyses the twitter-verse

Tips and tricks to speed up R and plotly based web apps

Benchmark of 5 popular graph/network packages - Networkx, igraph, graph-tool, Networkit and SNAP

Technical overview of our 2nd place solution and my experience at the Binance hackathon

In this post, I explore the problem of simplifying route intersections and document some Python code that can be used to clean and visualize Open Street Maps as a network representation

Part II in the network exploration of the Game of Thrones series. In this post, we combine the plots together and use gganimate to visualise relationships across all 5 books

A network exploration on the links between characters in the Game of Thrones series with the help of igraph and tidygraph

Chains, Forks, Colliders, paths and d-seperation - how DAGs can contribute to better causal inference

Deriving the OLS formula as a means of approximating the conditional expectation function

A reference cheatsheet on adjacency matrix, incidence matrix, laplacian matrix and the basics of algebraic graph theory

How should we choose the control group in a situation where we have multiple treatments and time periods? A simple statistical simulation exercise

Applying the SVD to the regression framework

To what extent do the coefficients obtained from a regression carried out at the group level correspond to the estimates at the individual level?

Deriving the OLS estimator via the maximum likelihood approach

A tutorial on using Leaflet in R for geospatial visualisation

Establishing the OLS formula via the method of moments approach

Deriving the OLS estimator - projection method

This post is the first in a series of my study notes on regression techniques. It covers regression as a solution to the least squares minimisation problem