linkedin · sam DOT j DOT coope AT gmail DOT com
DigitalGenius - Deep Learning Researcher
London: 2017 - Present
- Researching state of the art NLP - extractive question answering, word embeddings etc.
- Moving successful models from research to production systems.
- Developing fast thresholding algorithms to ensure highly accurate automations.
- Working with customers to improve model performance on their data.
- Teaching sales representatives on the AI platform, workshopping effective pitches.
- Interviewing and supervising new machine learning employees/interns.
- Refactoring the production codebase to a more functional/compositional structure.
Imperial College London - Computing MEng
London: 2013 — 2017
- Lead a group project to create a heartbeat sensor which uses video of the face and deep learning techniques. This was hosted on a server in a similar way to google deep dream.
- Created a back end service which authenticates users based on the way they type as part of a team web app project. Used PostgreSQL for managing the databases which stored users keystroke information.
- Worked in a team to create a compiler using ANTLR and java for a simple imperative language. Created an extension which used graph colouring to allocate registers.
- Completed PINTOS - a series of coursework aimed at improving the understanding of operating systems.
- Created an assembler and emulator for ARM architecture in a group project, then used the assembler to compile ARM assembly code into object code which activates lights on a raspberry pi when run.
- Completed weekly programming projects in Haskell, Java and C.
- Awarded a Computing Entrance Scholarship for outstanding A-level grades.
- Achieved a 1st honours for the second year of study including 91% in coursework and 92% in Compilers.
- Achieved a 1st honours for the third year of study.
San Fransisco/Mountain View: April — September 2016
Worked as a machine learning intern for 6 months.
- Developed my own 'medium-sized data' pipeline for use in machine learning.
- Used Go, react-redux for creating an internal tool.
- Used keras, scikit-learn and AWS for machine learning.
- Work was very independent, enjoyed the pragmatic culture of a startup trying to change the world.
London: July — September 2015
Worked in Core Monitors, a team creating realtime customisable spreadsheets for viewing market data. Developed in a proprietary Node.js like technology with Angular like front-end.
- Created the foundation for the mobile version of the new monitors app, moving a large client side application to a backend service, refactoring the codebase to allow for a UI-less version of the app to run on a server.
- Worked in a team following agile principles, and took part in formal code reviews for pull requests.
Birmingham: July — September 2014
- Created a tool which takes incremental slices of a 3D model, outputting the slices as models in the program, PNG files or SVG graphics.
- Wrote an API which, given a list of vector objects(beziers/shapes/straight lines), created a to-scale SVG file. This was then used for exporting vector graphics as part of a client’s request.
Bromsgrove: 2003 — 2013* Achieved A level results of A*AAA in Further Maths A2, Additional Further Maths AS, Physics A2 and Chemistry A2 respectively.
- Created a Google Chrome extension at HackKings ’14, ‘Baker’s Revenge’, which visualises cookies in your browser and allows you to delete them in a gamified fashion.
- Used Leap Motion and Tobii Rex at ICHack ’14 to create a unity game concept which used hand tracking on the Leap Motion and eye tracking of the Tobii Rex to navigate the user around a 3D space.
- Attended Nvidia deep learning hackathon where my team tried to create a network which could reverse engineer the latex from an image of a latex equation.