My experiments
Here you will find links to my experiments in my GitHub repositories - these samples are purely recreational and not made for production.
To find out about my professional experience, checkout my CV or simply contact me.
Reddit Bot with GPT2 and BERT
Fine-tuning a GAN-inspired generative LLM network with GPT2 and BERT.
Keywords: GPT, BERT, LLM, Fine-tuning
Stack Overflow Prediction
Training state-of-the-art AWD-LSTMs classifiers with fast.ai to predict Stack Overflow question answering.
Keywords: LSTM, fast.ai, Pytorch
SpaCy Resume Parser
A rules-based named entity recogniser developed to parse CVs for skills and score relevancy.
Keywords: SpaCy, Parsing, Named Entity Recognition
Landslide Detection
Using artificial intelligence to identify landslides in Hong Kong based on topographic, geological and rainfall data.
Keywords: CNN, Boosting, Blending
Chazbot
Using transformers to generate custom datasets and create a conversational Q&A AI.
Keywords: NLP, Transformers, GPT-3
Portfolio
This portfolio website was created using Django with a fully automated CI/CD pipeline for rollout.
Keywords: Django, K8S, Docker, CI/CD
Subscription Prediction
Binary classification on a skewed telemarketing dataset with emphasis on data exploration.
Keywords: EDA, Analysis, Bagging
Subscription Prediction
LSTM Autoencoder with Pytorch to detect anomolies in time series sequential data.