Santa Cruz, California, United States
I'm a programmer with a First-Class Honours degree in computer science and over 25 years of experience in the tech, finance, and health sectors. I'm a generalist who excels at greenfield projects and early-stage startups where flexibility is paramount. I'll be equally happy talking to your customers, coding the front and back end, defining your infrastructure as code, and doing on-call support. My career began in London with Cisco just as the Web exploded before I moved on to a telecoms startup and Yahoo! In 2004, I relocated to San Francisco, where I worked on a Web crawler for Alexa Internet, a wholly owned subsidiary of Amazon. Later, I worked supporting quantitative analysts and trading systems in the back office of two hedge funds before reinventing myself by starting Nourish Balance Thrive, an online clinic helping athletes overcome chronic health complaints and improve performance. I got back to coding after building a team of physicians, scientists, and coaches. I created bloodsmart.ai, a SaaS that uses machine learning to help clinicians make better decisions by predicting the results of blood tests. Seven years later, this business continues to operate without me. In 2022, I was the founding engineer at Engi, a VC-backed startup building a marketplace for software jobs, a bit like Uber for coders. Most recently, I've been working with a sleep scientist on SleepMate, an AI talk therapist designed to replace Cognitive Behavioural Therapy for Insomnia, the standard of care for insomnia. For the past ten years, I've hosted a health and fitness podcast with millions of downloads over 650 episodes. I'm a former elite-level mountain biker, expert snowboarder and kiteboarder, novice surfer and new student of Gracie jiu-jitsu.
Intelligent Document Processing, particularly in answering clinical criteria questions over unstructured documents using Gen AI: self-hosted LangSmith and LangGraph Platform on AWS with Terraform. Python.
Small Shopify store owners cannot do direct-response email marketing because they lack the time or copywriting skills. EmailWise helps solve this problem using large language models (LLMs). - React Remix, Javascript, Node, Python - LangChain & LangSmith (LLM monitoring and feedback) - REST, GraphQL, PostgreSQL, Redis - OpenAI, Anthropic, Meta AI, Docker
SleepMate is an AI assistant that uses large language models to replace Cognitive Behavioural Therapy for Insomnia (CBT-I). - Implemented evidence-based behaviour change techniques, including Motivational Interviewing and Acceptance and Commitment Therapy - Integrated data from validated surveys and wearable devices such as the WOOP wristband - Test-driven development practices ensure high-quality, reliable code - Built with Python, LangChain, Open AI Functions, Pinecone for Retrieval Assisted Generation (RAG) and MongoDB for long-term persistence - Discourse, Discord, and WhatsApp integration - Deepgram for speech-to-text and ElevenLabs for text-to-speech - Collaboration with Experts: Worked alongside renowned sleep physiology expert Greg Potter, PhD, and behaviour scientist and performance psychologist Simon Marshall, PhD
Engi is a VC-backed, open-source, decentralised software labour marketplace. Businesses define software jobs as a list of failing tests, and freelancers get paid automatically by making the tests pass. It's like Uber for coders. Working in a team of seven engineers, I used Test Driven Development (TDD) and Python to build: • The Engi marketplace command line interface to enable businesses to analyse their code and post jobs to a Substrate blockchain. The same CLI helped freelancers to solve jobs and get paid automatically. • The backend for a Figma plugin used by businesses to check if a React Web implementation matches the design. • The backend for a mass email REST server using Flask and AWS SES.
Dragon Rascals was an Ethereum web3/NFT fighting and racing play-to-earn game where I worked with three cofounders and three developers. • I wrote ERC-721 and ERC-1155 smart contracts in Solidity and successfully deployed them to the Ethereum mainnet. • Wrote Javascript to generate 8,888 unique images using hand-drawn layers. • Developed scripts to test, deploy, whitelist, withdraw, airdrop and pause smart contracts using Python and Brownie. • Built a backend server daemon in Python to listen to blockchain events and reveal metadata stored in AWS. • Wrote scripts to migrate ERC-721 metadata to IPFS, Pinata and arweave. • Built a web3/React/MetaMask minting app using Typescript framework useDapp. • Developed 2D games using Godot Engine.