London, England, United Kingdom
Data scientist and machine learning engineer with previous commercial experience as a software engineer in companies ranging from startups to large multinationals. Able to bring together experience from all fields to deliver impactful business solutions. Use Python and libraries such as Pandas, NumPy as well as Tableau, Kubeflow, Airflow, and NoSQL technologies. GitHub: https://github.com/tpgmartin Medium: https://medium.com/@tpgmartin Twitter: https://twitter.com/tpgmartin Newsletter: https://tpgmartin.substack.com/ Website http://tpgmartin.com/
• Created end-to-end cloud-based ML pipeline to enable clustering of over 50 million unique customers using big data technologies • Senior developer in internal project to scale out delivery of analytics projects, leading to 75% reduction in processing time • Led delivery of customer analytics report for a major business partnership
• Scoped and outlined data components of major B2B product, performing all data science work independently • Consultation and development of KPIs, metrics, and dashboarding for multiple initiatives • Development of novel model to predict time user email prompts • Ongoing support for data platform and MLOps
• Created probabilistic models for major public sector project • Curated data sources and produced multiple data components for B2B product • Created pre-employment screening POC using Ethereum blockchain • Created self-service Flask web app deployed on Microsoft Azure
• Initiated drive to become data-driven organisation, identifying KPIs with product teams. Creating visualisations and Tableau dashboards, integrations with Metabase and Slack. • Identifying customer segments using RFM analysis. • Identify trending StackOverflow topics using exponentially weighted moving average/variance of tag usage in new questions. • Creating data pipeline, reading from multiple third-party data sources and writing to AWS redshift. • Created Flask web app to view and summarise conference reviews. Ranks conferences using lower bound of the Wilson score confidence interval for a Bernoulli parameter. Use sentiment analysis to provide summaries of written reviews. • Mentoring junior members of staff on using data analysis tools. • Implemented features to user profile pages.