Neural Data Analysis:
Developed a pipeline to process ephys data from multichannel intracranial recordings, including spike
sorting and connectivity/causality analysis.
Started and coordinated a project to develop an optogenetic stimulation system small enough to be used in
an incubator. Working with the MIT Media Lab to develop algorithms to infer movement from EMG data.
Neural Network Theory:
Worked with IBM Research to develop an alternative to backpropagation through time (BPTT), solving the weight-symmetry and update locking challenges, enabling online learning in recurrent neural networks.
Cloud Infrastructure:
Created a system to optimize compute usage on the IBM Research cluster based on Ray.
I've built and manage my own small computation cluster with 10 CPU's.
(3D) Computer Vision:
Worked on a project using image segmentation and graph neural networks for indoor localisation.
Created a novel data acquisition method to train a prosthesis control system based on pose estimation.
Built a system to track and shoot down mosquitos. Worked on a project to segment and track living cells.
Robotics:
Lead a project to build a tendon driven robot monkey to study imitation learning in marmosets.
Built a robotic hand controlled with a glove and an automated device to shoot down mosquitos.