A lot of the projects that I end up involved in are fairly stealth. Here are some of the more public ones.


Beddy Butler

This open-source app is an experiment in Captology and Machine-driven behaviour change. It provides a gentle nudge to the user to go to bed at a reasonable hour (being an infovore, this is an issue that I suffer from).

Beddy Butler is a virtual butler in your ear who reminds you to go to bed, e.g. "It's getting rather late, your Grace", at a timeframe of your choice.

In fact, there are 3 sets of voices, from shy and retiring, to insistent and preachy, to a zombie butler (yes).

I did the butler voices myself (!), recorded with a special made-to-order Mitra 3D Mic Pro, specially imported from India, so that it sounds like someone is really over your shoulder. It was then heavily pitch-shifted to make my voice sound like an old man.

Props and sincere thanks to David Garces for profound help with the arcane and pointerless Swift, Livio Perissin for the awesome character graphics, Jeroen Muller for uncanny pitch shifting, and Abe Gong for the original inspiration, after I saw one of his talks at a Salon hosted by Alton Sun.

It might get it ported to Windows/*nix/Android if there's sufficient interest. It's a pretty hideous little app made just for my own ridiculous satisfaction, but the world has many diverse tastes.

Download it now off iTunes.

Or, why not fork it and record your own, as I've released it open-source (BSD license). 



EthicsNet follows the examples of WordNet and ImageNet to create a dataset of ethics for machine intelligence. EthicsNet is founded on the principle that ethics and morality are measurable, definable and computable, across cultures, time, and geography. We believe that the future of a thriving humanity in an artificially intelligent world is predicated on the ability to create moral and ethical machines.



An offshoot of the EthicsNet project, Bounder aims to enable a physical safety mechanism for autonomous systems such as drones. A small GSM-based tracking module can be attached to drones, so that they can be tracked in real time via web/app to ensure that they stay out of prohibited areas (or within bounds).

Also includes various pathfinding, maintenance, and mapping modules.



Scry was a prototype machine-driven health project launched during my study at SU that aimed to collect ophthalmological data from non-smart instruments using mobile phone cameras, and then pre-process it using cloud-based  Deep Learning / Convolutional Neural Networks, in order to pinpoint dozens of potential major health risks in moments.

It became sadly defunct due to distributed team inertia at the time, however these techniques are now being applied by Enlitic and DeepMind.