About me

Welcome to my website! My name is Josh, I’m a data scientist and software developer currently working at AWS.

My current role finds supporting AWS customers using SageMaker and Bedrock to do anything from training and deploying ML models to building custom LLM agents.

Prior to this, I obtained a Masters degree in data science and spent a few years at a YC startup.

Since I first began working in industry 8 years ago, I’ve found myself in entirely unfamiliar situations and been able to rise to the challenge every single time! For my next role, I’m looking to continue learning and challenging myself, and am particularly keen on roles that involve data science, AI, and building software.

In my spare time I frantically keep up with the latest developments in AI and enjoy reading about technology, psychology, and philosophy. See the bottom of this page for some examples of technical projects I’ve worked on in my own time.

I also have far too many hobbies; asides from the usual hiking and bouldering I am currently training for my first half-marathon, and DJ + produce electronic music.

Resumé / CV

Below is my CV. Also available as a pdf.

Work History

Cloud Support Engineer II (Big Data, AI / ML) - Amazon Web Services

January 2024 – Present

I’m currently working at AWS as a Cloud Support Engineer. My current role responsibilities are around building tooling and automation that improves the customer experience when interacting with AWS premium support. One of the projects I’m working on currently involves developing LLM agents for handling support cases and answer customer queries before a support engineer even gets involved.

This involves working within a team as a software engineer to write and deploy code to production, while leveraging my subject matter expertise with services like AWS Machine Learning services such as SageMaker, Bedrock, Forecast, Personalize, Textract, etc.

As a part of this role, I’ve also spent time directly handling support cases from customers. Given the nature of the services I am supporting this involves troubleshooting things like SageMaker training jobs/endpoints/pipelines, and Bedrock Knowledge Bases/Agents. This requires extensive knowledge about a wide range of topics: MLOps and the entire model lifecycle, LLMs, prompting.

I’ve also been involved in many other iniatives and value-adds such as: delivering training to other engineers, conducting interviews, formal mentoring, and project work.

Skills: MLOps and ML lifecycle, LLM agents, software development, customer support

Founding Engineer - CoffeeAI (Segna)

November 2021 – September 2023 (1 year 11 months)

I was part of the founding team of a Saas startup company. I followed the company through the prestigious Y-Combinator accelerator program and learnt best practice on how to build a successful startup, achieve product market fit, raise funds, and grow a company.

My main role was as a backend Python developer, but I covered many different roles and responsibilities across multiple different product pivots and iterations:

Backend Python developer: I was the primary developer of the flask backend written in Python. I wrote and reviewed code, architectured the backend, and worked with the frontend team to design and manage APIs.
Data scientist: I implemented machine learning models to learn on user input and improve the user experience. I was also responsible for visualising and analysing usage metrics to inform product/marketing decisions.
Prompt engineer: I worked directly with the OpenAI and Anthropic APIs to build a generative-AI powered application. I was responsible for designing and testing the prompts used to generate the AI output.
Cloud Infrastructure / Devops: I was responsible for managing the cloud infrastructure on GCP and CI/CD pipelines on Github Actions. I configured Kubernetes for both a stateful and stateless application. I also managed and helped to design the PostgreSQL database.
Content marketer: I spent some time generating marketing content for the company in the form of TikTok videos and blog posts.

Overall, I was relied upon to reliably be able to solve difficult engineering problems. I worked both independently and as part of a development team and managed interns. I was an invaluable member of a small team and really enjoyed having a direct impact on the direction and success of the company.

Skills: running a startup, backend software development and architecture, database design and administration, cloud deployment and devops, prompt engineering

Masters Student - Luma Analytics

July 2021 – November 2021 (5 month dissertation)

For my Masters dissertation I worked with a data analytics company. I used data to improve the analytics and insights available for amateur golfers, and uncover trends in the sport at large. This involved PCA, clustering, data visualisation, web scraping.

The work was highly exploratory and consisted of self-directed research in an unfamiliar field with only low-touch supervision.

I received an A+ grade and my work was highly valuable to Luma. It went on to be continued by another student the following year. You can view my dissertation here.

Skills: principal component analysis, clustering, data visualisation, web scraping, technical writing

Computer Vision Engineer - Beca

November 2020 – February 2021 (4 month internship)

I worked in a team to develop an end-to-end computer vision solution encompassing data labelling, model training, and prediction validation. I researched contemporary computer vision models before implementing Faster-RCNN and YOLOv5 in PyTorch, tuning them to the dataset, and comparing their performance.

For this role I acted as Product Owner and team leader of a group of three interns. The project had minimal supervision—we met with our supervisors once per week and were responsible for coming up with a sprint goal each week and delivering it. As team leader I oversaw planning and delegation of tasks, while also being heavily involved in the technical work.

Skills: computer vision, agile software development, dataset curation and labelling