Long term projects
Consider the variety of ways humanity stores its knowledge: the medium, structure, language, and means of access. With the extreme variety comes silos and disconnects. It is likely we know more than we think we do, the information is just ininaccessible to the ones who need it. I want to explore and capture ‘knowledge’ in the most abstract sense and make it generally accessible, to both humans and computers.
Increasing pressure on peoples time, nutrition, and food waste are requiring higher levels of automation. I want to capture food preparation processes and use them to guide the design of a fully automated kitchen. This kitchen would be able to prepare any known recipe. Lessons learned from this project will feed back into the Knowledge Capture project.
Data science is rife with many disparate resources, data repositories, tools, and standards. I want to unify the various resources associated with data science. This involves fully capturing the tools and data in a way that respects the practical implications of the supporting infrastructure. Lessons learned from this project will feed back into the Knowledge Capture project.
“Blindness cuts us from things, but deafness cuts us off from people.” -Helen Keller
Coping with hearing impairment comes with numerous challenges and depending on the available support systems can lead to lifelong troubles functioning in society. I have identified some readily available technologies that when combined will bridge the education and communication gap between these peoples and their communities.
Projects backburnered due to time, resource, or demand constraints
I wanted a quick and portable way to deploy positional audio into a space. This project explores using Bluetooth speakers and PulseAudio to accomplish this.
The ESP8266 is a very cool little WIFI enabled uC and the market is flooded with cheap modules that have been built around it. I am looking at various solutions that can be built around it to make renter friendly smart home devices.
Most python libraries that allow reading and writing HTS data (sam/bam) either link against HTSLib externally or are limited in functionality. I needed a python native implementation of HTSLib to allow compatibility with the Numba project. This will allow for much faster performing python code and even GPU accelerated algorithms.
I was not satisfied with the command line argument parsers available for python and wrote my own. It was built to minimise the amount of code needed while maximising flexibility. This is intended for applications where the command line arguments are numerous.
Ever see those blinds that automatically go up and down in smart homes? Well I needed a more renter friendly solution. There are some good ones on the market but they are over $100 per window. I know I can build it cheaper.