danrhowarth Goto Github PK

I am a data science leader who has successfully delivered consultancy and platform focussed data science products. I am at my best working with customers and other stakeholders, to help define a problem and develop and deliver ML-based solutions to solve that problem.

My LinkedIn profile is here, and a copy of my CV is available here

My professional experience is focused on the industrial sector, developing and deploying anomaly detection algorithms for use cases such as predictive maintenance or network security. This includes Deep Learning and more traditional machine learning approaches, and cloud and edge based deployments.

I am a strong technical communicator and have taught data science at the University of Oxford, and written blogs for Gantry, a cutting-edge MLOps-focussed Silicon Valley start-up.

I work hard to keep my skills relevant: a recent focus has been MLOps and state-of-the-art Deep Learning.

This repo includes the following:

• An example of a robust, reproducible ML pipeline using Weights and Biases for experiment tracking
• Deploying ML Models with CI/CD, using FastAPI, Github Actions and Heroku
• Model Monitoring and Scoring using Flask

• Deployment of Image Anonymizer application using Streamlit, Docker and drawing on state-of-the-art models (the output of a Full Stack Deep Learning project group I was part of, and which was selected as a top project for the 2022 cohort)
• Deployment of Text Recognizer, with inference performed via severless GPUs using modal (this repo represents the teaching material for the Full Stack Deep Learning course, and I then subsequenty contributed to the severless GPU implementation)

• A repo containing the notebooks and other teaching material I developed for the University of Oxford is here
• And some tensorflow2 notebooks I developed for a series of workshops are in this repo
• The blogs I have written for Gantry are available here, covering MLOps, LLMs and other topics

• A repo that implements a personal Computer Vision project and which was presented at PyData Bristol
• Notebooks and other material from ML Dev Ops Nanodegree, and ML Engineer Nanodegree