Experience

• As a research engineer I was involved with developing novel proof of concepts for a variety of stakeholders in both the public and private sectors. As well as technical work I have successfully project managed several projects and contributed on successful project bids. In the cultural heritage sector, I helped to build a content aware archiving solution based on a novel chunking technology. This later became a spin-off company. In the security and surveillance sector, I built a number of camera systems for rapid deployment in different environments such as car manufacturing plants, airports, and water plants. Additionally, I was responsible for taking research algorithms and making them work real-time in-situ. This included feature extraction, tracking, and high level reasoning with data from a range of novel sensors (vibration, magnetic, etc). In the educational sector, I worked on a platform for writing prosocial games. Specifically, building voice emotion classifiers and a data fusion pipeline. In web analysis, I worked in a number of projects including analysing twitter data (location extraction), sentiment analysis, and fast internet search and indexing. Much of this work involved the use of a low-level FPGA card that performs fast regular expressions. An appliance was built to aid the use of the card which has been exploited by partners as the basis of a startup. In addition I am on the systems team and am specifically responsible for our Gitlab and Gitlab-CI. Furthermore, I have been deeply involved in the company's containerization strategy. Key technologies: * Devops: Vagrant, Virtualbox, Ansible, Gitlab * Linux containers: LXC/LXD, Docker * Analysis: Numpy, Pandas, OpenCV, MATLAB * Machine learning: scikit-learn, MATLAB * Programming: Python, C/C++ * Web frameworks: Django, Flask, VueJS * Visualisation: D3JS, Jupyter, Matplotlib, Bokeh, MATLAB Show less



• Designed and built an autonomous, automated biometric capture environment for the UK Defence Technology Centre * Infrastructure development : Installed and maintained a cluster of 10 Linux computers with Gigabit networking. This included a dedicated file store. * Camera control : Wrote C++ code to abstract a Firewire camera driver. Made camera access and control easier. * Software Middleware : Wrote a lightweight agent framework providing a directory service and a communication system. * Camera Calibration : System to automatically calibrate an array of distributed cameras. Employed edge detection, Hough transform, and colour image processing techniques. * Segmentation : Wrote code to segment a moving object from a static background. Technique based on image differencing and shadow suppression in HSV space. * Face extraction : Designed system to segment only the facial region from image sequence. Exploits anatomical properties of face. * Code optimisation : To speed up code sub-sampling and algorithmic assumptions were employed. 3D reconstruction : Employed shape from silhouette algorithm to produce a 3D voxel representation of a subject. Data comes from 8 cameras distributed around the network * Storage : Code to save the image/3D data to the file store. * Scalability : Used a XML data catalog for stored data. Allowed scalable data collection and also provided an audit trail for processing * Project management : Low cost footfall sensor, 3D ear scanner, Extracting 3D human motion from a single camera view. Show less