Who is Andrew Fraser?

Good Question!

I am a 27-year-old programmer with nearly 20 years’ worth of experience being a massive nerd. I currently work at AND Digital as a Senior Product Developer working for a variety of clients. Delivering high-quality frontend applications and websites, helping clients improve their Agile artefacts and trying to upskill all members of the teams.

Before that I worked for High Speed Training in their Learner Engagement team; focusing on collecting user engagement using Azure Application Insights, creating templates for our in-house content authors to use in our Angular Single Page Web App, expanding our RESTful API functionality and maintaining excellent Lighthouse scores for all of our pages.

What do you do…generally?

In my spare time, I enjoy automating various things around my house, reading books outside and getting a bit too competitive at pub quizzes. I live with my partner Charley and we’ve yet to have any pets. But when we do buy a house we want to buy some chickens that we have already named Jelly, Jammy and Gemma. 

What do you do…here?

This blog is home to a lot of blog posts. Those that cover various tech trivia and gadgets, namely smart home devices. As well as discussing various professional ideas of which opinions are solely my own.

At work I specialise in creating extremely high-quality, TDD, projects that are extensible, accessible and versatile and outside of work I specialise in Pub Quizzes, running 5k’s and finding bargains online. In the past, I have focussed on areas of, as the government is now calling it “Cyber”, such as Digital Footprints, Neural Networks and even straying into iOS app development for the National Railway Museum.


I’ve worked with a variety of clients at AND Digital, some of whom I can’t disclose on this blog. 
But I can say that I’ve worked on Extremely large-scale websites for very well-known companies. Delivering very performant work, in a timely manner to the highest of standards. Quickly moving up the ranks within the company receiving some very lovely feedback as to my efforts. 

Digital Footprints

In 2016 I did a short 15-minute presentation on Digital Footprints and what they really mean when you get rid of all the technology mumbo-jumbo.

I demonstrated I could acquire extensive personal information about a subject. Their family members, finances, political views and had a great footing to start Social Engineering this person. Ending the presentation with tips on how to secure yourself online and a Q+A that delved into the more complex nature of securing friends and family online.

Neural Networks

My Dissertation at University was titled, “Sentiment Analysis on Shakespeare Sonnets written using Andrej Karpathy’s Recurrent Neural Network.”  It focussed on rehashing a Neural Network and feeding it all of Shakespeare’s Sonnets, getting the results from the Neural Network after various iterations. The results from this were then put through spellcheck and then handed to a group of testers and asked to be rated on both content and emotional impact.

The conclusion?

The original hypothesis stated that if a Recurrent Neural Network could replicate something that could be mistaken for an Elizabethan Sonnet, specifically a work by Shakespeare, then the project would be a success; with ‘replication’ defined by promising results after sentiment analysis.

The original target, as illustrated in Chapter 4, Section 3, was to create a recurrent neural network with around 89.6% accuracy, similar to that of Shakespeare, (Words he used, versus words he created). In the Sonnet seen in Chapter 5, Section 1, Part 6, which was reached in about 30 minutes, 83 out of the 100 words produced were valid or pseudo-valid (following English grammar rules). This figure is only 6.6% behind the yardstick set out at the outset for Shakespeare’s accuracy: perhaps the greatest writer in history, with a lifetime of linguistic experience and human emotion to draw upon. So while the RNN-written sonnet hardly qualifies as great literature, it could be argued that as the Recurrent Neural Network started with no knowledge of the English Language – and had, presumably, never had its heart broken – to create the sonnet that it generated after just 30 minutes is a highly promising result.

iOS Development?

Yes. iOS Development.
I spent a term in my final year at University, between classes, working with the lovely folk at the National Railway Museum (NRM), in making one of their extremely interesting data collections, accessible to a wider audience.

The app titled, “Fallen Railwaymen”, was an iOS application, that was responsive on all iOS devices, that made available a database of 11,000 strong World War soldiers that had also worked for the railway, but fallen in battle, and some location, trivia and rank of the officer.

Splitting the app up to using an Apple Maps API, to pin the locations of the soldiers and their homes, on a map to investigate.
A quiz on other War facts related to both the data and the railway in general.

As well as an efficient search functionality to allow people to search soldiers by location, surname, rank or battalion.