I have led or consulted on R&D and product development efforts in a wide variety of industries, from petrochemical to environmental to biomedical. Some of the solutions that I was responsible for include:
Intelligent web app for diagnostic strip reading. Using a custom-compiled wasm version of openCV and a custom-trained neural net, this app provides a biomedical start-up with a user-friendly way to improve the precision and value of their diagnostic test kits.
AR-assisted thin-film measurement web app. I developed algorithms and built a progressive web app to read diagnostic components for a medical nanotech customer using marker-based augmented reality, mobile device sensors (camera, GPS, accelerometer), and a distributed cloud database.
Industrial internet-of-things oilfield chemical injection monitoring system. I managed a product redevelopment program, including market research, pre-selling, and product redesign to fit market needs. The redesign included new embedded software, IOT connectivity leveraging Particle.io and Google Cloud services, a web app for configuration and control, all with a 3-fold reduction in manufacturing costs.
Chemical product introduction and market development. To solve major customer issues for a major specialty chemical company, I worked with their sales force and internal R&D group to introduce 3 new products which now account for 7-figure annual revenues. From preventing a $120 million outage with an antifoulant to process modeling to drive new corrosion inhibitor product development, my data-driven approach brought engineers, scientists and management to rapid agreement and efficient implementation of solutions.
Diagnosing unexpected water spikes in a desalter. I performed 4D CFD to verify my initial opinion that a component design was setting up adverse transient flows in the unit, working against the separation process. The customer changed out the component and obtained stable performance.
Neural Net model of an oilsands extraction train. Using careful data pre-processing based on plant experience, standard machine learning methods were trained on historical plant data to predict product quality and identify deviations in real time.
Using a dispersant to dislodge fouling in a fractionator overhead. With fouling trends predicting less than 2 weeks before a very expensive plant shutdown would be required, I recommended and implemented a salt dispersant chemical program to move the deposits to where they could be solubilized and removed. The unit responded within days, and the program has been in continuous use since. The program expanded to other units and accounts for 7-figure annual revenues.
Record-setting bioremediation of creosote contaminated soil. A small start-up contracted me to help optimize an enclosed soil bio-remediation system. By finding unique ways to optimize parameters, we achieved an 85% reduction in heavy polyaromatics in 7 weeks.
Deep-relief chemical milling system for marble tiles. A local stoneworking firm wanted a cost-effective way to create sculptured marble tiles. I invented, built and programmed an automated chemical milling system that could mass-produce deeply sculptured patterns and figures in limestone or marble slabs or tiles.