Work

Making data science work for the financial industry at GreenKey Technologies

Voice transcription unlocks hidden value in the financial industry. Helping traders record what they say - and when - isn't just good for compliance, it opens the door for internal analytics and smarter decision making. To feel the pulse of a voice-driven industry, you need to capture key patterns in large amounts of audio. At GreenKey Technologies, I employ state of the art machine learning tools to extract information from real time voice streams.

Past work: Harnessing data science to make clean energy

Emerging chemical materials can capture abundant solar energy and use it to power our lives and keep our environment clean. Computer modeling promises to save countless man-hours of synthesis and get these products to consumers faster. However, existing approaches cannot readily predict what chemicals should be developed. I designed new methods for describing structures and properties, leveraging the power of modern computer architectures to balance accuracy and cost, employing machine learning techniques to elucidate structure-property relationships, empowered by state-of-the-art big data toolsets.

The performance of organic photovoltaic materials depends on electronic transport and structure at several different lengthscales. We established how the conformation of organic photovoltaic polymers changes the transport properties in order to understand the performance enhancements reported for ternary blends. Our coupled classical and ab initio molecular dynamics simulations showed that polymer twisting reduces optical absorption efficiency, as well as hole transport rates in donor polymers. Using these results, we showed that hole transfer rates are enhanced in ternary blends due to morphological and energetic changes which occur synergistically.
"Planarity and multiple components promote organic photovoltaic efficiency by improving electronic transport ", Matthew Goldey, Daniel Reid, Juan J. de Pablo, and Giulia Galli, Phys. Chem. Chem. Phys., Advance Article (2016)

About

Matthew is a data scientist at GreenKey Technologies unlocking the power of voice transcription for stock brokers and traders.

CV

-->

Elements

Text

This is bold and this is strong. This is italic and this is emphasized. This is superscript text and this is subscript text. This is underlined and this is code: for (;;) { ... }. Finally, this is a link.

Heading Level 2

Heading Level 3

Heading Level 4

Heading Level 5
Heading Level 6

Blockquote

Fringilla nisl. Donec accumsan interdum nisi, quis tincidunt felis sagittis eget tempus euismod. Vestibulum ante ipsum primis in faucibus vestibulum. Blandit adipiscing eu felis iaculis volutpat ac adipiscing accumsan faucibus. Vestibulum ante ipsum primis in faucibus lorem ipsum dolor sit amet nullam adipiscing eu felis.

Preformatted

i = 0;

while (!deck.isInOrder()) {
	print 'Iteration ' + i;
	deck.shuffle();
	i++;
}

print 'It took ' + i + ' iterations to sort the deck.';

Prototyping molecular dynamics in python

Matthew B. Goldey

From my graduate work at Berkeley, I developed a love of weak interactions between molecules and testing methods in Python. The strength of weak interactions (as compared against the kinetic energy from heat) determines whether a bunch of molecules will exist as a gas or as a liquid. [Un?]fortunately for theoretical chemists, the energy can be understood using formulations like the Lennard-Jones potential $$ V_\mathrm{LJ} = 4\varepsilon \left[ \left(\frac{\sigma}{r}\right)^{12} - \left(\frac{\sigma}{r}\right)^{6} \right] = \varepsilon \left[ \left(\frac{r_\mathrm{m}}{r}\right)^{12} - 2\left(\frac{r_\mathrm{m}}{r}\right)^{6} \right], $$ which works well for rare-gas based materials and coarse-grained models. However, even if we understand what the interaction is between particles (molecules, rigid rods, cats, etc.), we need to have a system of integrating these equations with respect to time. One good method for integrating equations of motion is Velocity Verlet, which has a comparatively small error as a function of time-step. Given an interaction potential, a system for propagating position and speed, we can map out how interaction strengths result in mixing or phase separation between materials. Python is an excellent tool for prototyping how time-steps, potential shapes, and thermostats result in different dynamics. Below is a matplotlib animation (h/t Pythonic Perambulations) which shows the mixing between phase-separated materials as a function of time. At high temperatures, the mixing proceeds smoothly.
For lower temperatures, the mixing occurs much more slowly, generating a rough phase boundary which appears to persist.

Code available here