Selected work
21 clients. Six highlights. One approach: go deep, move fast, and don't stop until it works.
Rebuilt their entire performance marketing stack from the ground up. New acquisition funnels, A/B testing infrastructure, and a data pipeline that let the team make decisions in real time instead of two weeks later.
$15M revenue growthBuilt automation to track and optimise ROI across campaigns, removing manual work and keeping performance aligned.
~121 hours saved per monthBuilt a Facebook competitive-intel and optimisation system that became a core lever for scaling acquisition.
300+ hours saved per month · helped drive $15M revenue growthDesigned and built an AI recommendation engine for government health policy. Processes thousands of policy documents and surfaces prioritised, evidence-backed recommendations to decision-makers.
Policy engine at scaleBuilt and maintained the consumer-facing web application, giving B2C users a clear pathway to value and engagement.
Created and maintained the business application for B2B clients—helping them identify revenue opportunities and map pathways between B2B and B2C.
Improved and maintained their Flutter mobile application, keeping the product aligned across web and mobile.
Shipped a full AI SaaS product in under a month. From architecture to UI, from auth to billing. The kind of build that usually takes a team of four six months. We did it in four weeks.
0 → 1 in 28 daysConsulting and development on Discord bots—strategy, architecture, and implementation.
Designed and built agentic AI bots for automated workflows and user-facing interactions.
Built a crypto aggregator with leverage trading across 2000+ symbols, support for 4+ chains, and integrations to six DEX platforms.
Heavy focus on geospatial data: scraping 60 million data points live every three days, turning raw web data into structured JSON that made sense and was used to pitch projects. Built pipelines and models that turned disparate sources into actionable intelligence for policy and investment.
60M data points · live every 3 daysMapped every manual workflow across their operations team and automated the ones that were killing productivity. Built on n8n, custom Python, and LLM-powered decision nodes.
642 hours per month saved