Difference between revisions of "CV:Publish"

I lead an ongoing research effort to precisely quantify power loss from damaged or old photovoltaic panels in commercial solar power plants. Infrared images of the panels under electroluminescence are collected by a drone-mounted InGaAs detector, and pixel-wise intensities are used to fit an equivalent circuit model, which in turn approximates mismatch loss and I-V characteristic.

• Designed and implemented a distributed messaging framework to support pilot-from-shore capabilities for underwater vehicles.
• Developed a controller to distribute energy from a hydraulic powerplant in a tractor-drill combine depending on workload or sensor input.
• Prototyped a Pytorch-based computer vision system for robotic survey of undersea pipelines.

Completed a feasibility study on sentiment analysis of images in social media, funded by a research grant from the UK government's technology strategy board. Starting from the Yfcc100m and YLI datasets comprised of 100 million images, labels, and metadata, I investigated both novel and existing methods and developed a commercial product, which has since evolved to be based on convolutional neural nets.

Completed a research internship through the Bright Minds Intern Competition programme in the Machine Learning and Perception research group, working with Principal / Senior Researchers Drs. Pushmeet Kohli, Yoram Bachrach, Ulrich Paquet, and Filip Radlinski.

I worked on Project SmartFence - an application for web access control. Users block or allow a few sites they know about, and SmartFence automatically infers the suitability of the rest of the web. We developed several different cluster/kernel-based models and visualization schemes. The final model generates a high dimensional embedding of websites from search sessions (think associated filtering). I delivered a prototype for the OneWeek company-wide hackathon, and a patent was applied for.

Internship with UniEntry to develop a pilot site to help sixth form students find the right university. Developed a platform that filters information from the UK's Higher Education Statistics Agency and gives recommendations based on students' registered information and grades.

Competition to implement a performant implied volatility calculator. Team awarded second-best for code review and performance, and best for presentation.

• AA274A Principles of Robot Autonomy I (Fall 2020) GPA 4.0
• CS234 Reinforcement Learning (Winter 2019) GPA 4.0
• CS231n Convolutional Neural Networks for Visual Recognition (Spring 2017) GPA 3.7 - Project: Bounding out-of-sample objects

• Final year research project - Predicting Personality from Twitter
• Information Retrieval and Data Mining Prize - awarded for research and public poster session
• Best Undergraduate Research Group Project of the Year - Task Identification Using Search Engine Query Logs
• Developed an Android app for the Restless Beings charity to conduct field studies on children in poorly-developed countries

• Withdrew in second year to transition to computer science.

• Outstanding Student of the Year 2008 - Double award (Chemistry, Music) | Most imaginative hovercraft design