Project Highlights

SortStream - ML powered document sorting

Developed an application to automate the sorting of documents based on explainable natural language processing. This could be deployed without technical expertise and greatly reduce manual effort. Now available via pip install sortstream.

DocDump - NLP Python Package

DocDump aims to allow for raw text data and document metadata to be easily extracted from a range of commonly used document types. Now available via pip install docdump

Service deserts in Victoria

Using the google Maps API and poisson regression to perform a statistical analysis of food, banking, health and education accessibility in Victoria and how this accessibility varies with local income.

Automated Data Exploration

Well presented data exploration reports created with R, RMarkdown, ggplot. Used for the initial analysis of a dataset and to inform the next steps in analysis or modelling.

Look and Learn - Data Science Proof of Concept

Lead a proof of concept project to develop an unsupervised image embedding and retrieval system to provide image recommendations for marketing purposes as well as aid in image sales analysis. Look and Learn is the world’s largest and most unusual collections of historical pictures and vintage imagery, with half a million illustrations from original artwork, books, magazines, illustrated papers, posters, prints, postcards, trade cards, old photographs and ephemera of all types.

Other Projects

Navigation App for Urban Cyclists

Developed an iOS app to experiment with alternatives to turn-by-turn navigation. Developed in Swift

Exploration of Document Classification Techniques

Classifying news articles wth CNN, Dense, and Logistic models

Deep Dream Implementation in Keras

Visualising CNN learning and creating some neat art

Viral Fashion Simulation, Random Graphs & Statistical Physics

Simulating why Germans gave up ties with an interacting agent approach

Market Evaluation with similarity methods

Exploring how indicator similarity can be used to understand markets and economies

URL Sketchifier

Give your urls a look that Gmail will flag as suspicious

Logistic Regression — Understanding Explainable AI

Implementing maximum likelihood estimation for binary logistic regression

Decision Trees — Understanding Explainable AI

Learning how decision tree algorithms work

AI, Connect 4, and finally being good at this game

I built a mini-max algorithm to beat my friends, and embedded it in a Facebook messenger bot

About me

I’m currently working at Deloitte Australia in their Analytics and AI team, after graduating from a Bachelor of Commerce in Economics and Finance at the University of Melbourne.

After growing up in South Africa, I’ve lived in New Zealand and now Melbourne, Australia. I’ve always enjoyed tinkering in a workshop, and in the absence of power tools this has morphed into tinkering with programming, ML, and electronics. This has spurred on many side projects, some of which are documented above.

Education and Experience

Things I've done or am still doing

Consultant at Deloitte

Working in the Analytics and AI department of Deloitte Consulting, Melbourne.

Prior positions: Graduate, Analyst

(2019 - Present)

Bachelor of Commerce, Univeristy of Melbourne

Graduated top of the faculty after studying Finance and Economics, while distracting myself with a minor focus of Structural Engineering and Econometrics.

(2016-2018, First Class Honours, Faculty Medal Recipient)

Academic Tutor, Queen’s College

I have been tutoring undergraduate Macroeconomics, Microeconomics and Econometrics, as well as masters Econometrics. Previously I taught IB economics.

(2016 - 2020)

Data Scientist at Look and Learn

Lead a proof of concept project to develop a search engine that captures art style, as well as a project to automate keyword generation.

(2019-2020)

Treasurer for The Wyvern Society

Acting treasurer for The Wyvern Society, the alumni society for Queens College.

(2019-Present)

Volunteer at Big Brothers Big Sisters

BBBS is a volunteer youth mentoring service, aiming to provide positive role models to young people.

(2020-Present)

Intern at Deloitte Consulting

Worked in the Performance | Automate department of Deloitte Consulting on data sourcing for the database migration in a large merger.

(2018)

Intern at Viburnum Funds

Viburnum Funds is an Australian PE and strategic equities fund. I worked in the research department, doing deep research into potential investment opportunities. Gained experience in modelling high growth companies and in industry research.

(2017)

Writer at Towards Data Science, Hackernoon

Hackernoon and Towards Data Science are the 4th and 13th largest online publications on Medium.com. I write on a range of technical subjects, with my stories gaining over 36,000 views.

(2017 - 2018)

Business Analyst, CSBA

Working in the consulting unit of CSBA. Provided a review of internal processes and products and research for engagements with clients.

(2017 - 2018)

Co-Founder, Line2

Started Line2 with the aim of providing passive, low cost customer insights to small retail and hospitality businesses. We didn’t succeed, but I did learn a great deal about data management and analytics, privacy, and hardware development.

(2017)

International Baccalaureate, Kristin School

Completed IB in 2015, taking English, Spanish, Economics, Chemistry, Mathematics and Physics.

(2014-2015, 44/45)

Get in touch

Feel free to send me an email on gwholtes@gmail.com or message me through LinkedIn.