Sentence Embeddings and CoreNLP's Sentiment Model
Published in Towards Data Science
Hi, I am Laura.
Welcome to my data science portfolio!
Laura in a nutshell:
UCL Postgrad
Natural Language Processing
Social Network Analysis
Neural Networks
Raspberry Pi
Literature and Linguistics
Data cleaning & Visualization
Databases and SQL
About me
Hi, I am Laura! I currently work as a Machine Learning Engineer at Spoon Guru, where I build end-to-end Machine Learning pipelines. At SpoonGuru I am able to apply my Machine Learning expertise to grocery data and build data products for the retail sector. Before SpoonGuru, I worked at Codec.ai. Where I specialised in behavioural analysis of users through the study of text using Natural Language Processing (NLP) methodologies.
I have a very non-conventional, interdisciplinary background. As a bachelors, I studied a BASc in Arts and Sciences in University College London (UCL). This is a interdisciplinary degree that allowed me to freely combine my passion in literature with modules from other areas of knowledge. I majored in literary and linguistic studies and minored in computer science. It was during this course when I discovered and fell in love with NLP, which I explored in order to make a computational literary analysis of Walt Wiltman's Leaves of Grass for my final year dissertation.
After BASc, I graduated with distinction on a MSc from the Information Sciences department of UCL. This course allowed me to further explore the field of NLP and introduced me into ML methodologies. Both of these worlds intersected in my dissertation under the title of: Cross-Domain Evaluation of Document Sentiment Classification using a Combined Recursive-Recurrent Model. My expertise on the field of ML allowed me to work as a Research Assistant at UCL for Dr. Luke Dickens and help him create teaching material for the module Introduction to Machine Learning and Data Science.
My NLP publications:
Intro to Stanford's CoreNLP for Pythoners
Published in Towards Data Science
Transfer Learning and RNN — the CoreNLP Series
Published in Towards Data Science
Other NLP projects:
The Electionary: Textual Analysis of U.S. Presidential Elections
The Electionary is a project that analyses historical trends in the use of language in American politics, more specifically the language used in Presidential election campaigns. We investigated the ways in which use of language has changed, by using computer-aided textual analysis methods primarily on the text of the television debates in the campaigns from 1960 onwards, using the American Presidency Project as a resource.
My Raspberry Pi Tutorials:
Raspberry Pi 3: connect to multiple WiFis & set multiple Static IPs
Published in The Startup
Raspberry Pi 3: headless setup with Mac
Build a Mobile App and trigger Subprocess on your Raspberry Pi
WebApp to send Message to Raspberry Pi Screen
Courses and certificates
Mathematics for ML: Linear Algebra
Deep Learning and Neural Networks
Inferential Statistics
Structuring ML Projects
Production ML Systems
Deploy ML Models to AWS Cloud Servers