Projects +

1. Pyspark analytics tool

   A python application that gathers data from the client's businesses analysies it and exports it in various ways, to give flexibility to the end user. The application uses spark and hadoop, submitting jobs using yarn.

   In this project I wrote spark queries in order to transform and aggragate data. I also developed a unit testing framework that creates an API to assist with the creation of unit tests for the spark queries and offers a unified way of creating spark dataframes and asserting the equality of dataframes in a seemles way. I used Freemarker Template to generate configuration files in order to enable the adaptation of the software from other departments, without the need for futher development.

   Configuring and fine-tuning spark was one of the most interesting parts of the project, as the documentation for doing so merely covers the simplest use cases. Another interesting aspect of the project was writting unit tests around spark prudently because of the overhead imposed by the creation of a spark session. It was crucial that we only included sensible unit tests so that the build times wouldn't grow in a disproportionate manner.

2. Green New Deal Glasgow (Common Wealth)

   A series of interactive websites that present the potential implementation of the Green New Deal in Glasgow. Research that was conducted by Ground Lab and presented by the Common Wealth’s Green New Deal series. Sample

   

3. UK Restoration and Rewilding (Common Wealth)

   An interactive SVG map that represents the benefits of re-wilding the UK based on a quantitative research by Simon Lewis as part of Common Wealth’s Green New Deal series. Sample

   

4. Green New Deal - City of the future (Common Wealth)

   An interactive SVG map that depicts an implementation of the Green New Deal. Sample

   

5. Fscs Reporting tool

   A Java application that gathers data from the client's businesses and produces a report that is compliant with the FSCS regulations. In this project I promoted the use of Java Stream, in order to express the many complicated transformations of the data flow.
   Details cannot be provided due to a non-disclosure agreement with the client.

6. Candidate Application Tracker

   A web app that tracks the progress of a candidate for any of the offered positions. Part of the project was to maintain and add features to the Java back-end, while updating the Durandal front-end. Another part was to entirely re-write the front-end with ReactJs while making it GDPR compliant.

   This application is integrated with Exchange in order to book meeting rooms and invite the potential interviewers. It is backed by an SQL database in order to keep interview scores, comments and help the HR team select the most adequate interviewers from the company staff, according to experience and personal involvement with relevant technologies to the position. Involved heavy use of granular user privileges according to the user's position, level of seniority. It is integrated with the company website in order to achieve an automated application process with the wpengine and Gravity Forms.

7. Website of Artist (University of Bristol)

   This project began with the Web Technologies unit and was later developed into a collaboration with a real artist. This project received best of class mark. Sample

   

8. Web based real estate speculation game participate

   Live-prototype of an entry for the Tomorrows exhibition, organised by the Onassis Cultural Centre.

   

9. University forum website database (University of Bristol)

   This was a group project (4 people) for the databases unit. Each member of the group was responsible for the implementation of separate method, each one of which included an SQL query and the equivalent Java server method. Key part of the project's success were the productivity platforms and the test driven development. This project recieved best of class mark (100%).

10. Mammography Interpretation Tool
    using Computer Vision (University of Bristol Dissertation)

    Convolutional Neural Networks (CNNs) are machine learning structures able to implement aspects of biological vision systems. They can be used in Computer Vision applications to recognise and classify objects in given imagery.
    
    This project aims at producing a mammary gland screening tool using CNNs. A tool that is, at least, able to segment volumes with densities that differ from the expected in a provided mammography sample. The project aims at highlighting suspected regions of malignant tumour in these segments. The system would ideally be able to assist doctors in the assessment of mammography samples.

    
    

    

    

    The framework used for the CNNs is Caffe, a framework developed by Berkeley Vision and Learning Center.

11. Experimental Corroboration of
    Reduced Beam Section moment connections (DUTh Dissertation)

    The aim of this project is to protect connections and columns in steel frames, from brittle failure, by the controlled weakening of the beam. This project included a full scale Eurocode compliant epxeriment and its corroboration from a Finite Element Analysis model run in Abaqus/CAE.