Current Roles

Senior Data Scientist — Technical Lead, Climate & Global Health

Office for National Statistics — May 2024 to present

Technical lead for SOSCHI (Standards for Official Statistics on Climate-Health Interactions), a Wellcome Trust-funded programme hosted on the UN Global Platform, and ONS technical lead for BREATHE.

  • Lead the team that built the climatehealth R library and its Plumber API, so a single tested implementation serves a Flask web platform, an R Shiny dashboard and command-line users. I set the design and engineering standards and matrix-managed the developers who wrote the functions and modules, coding myself where the team was stretched or the work needed it. Past 2,000 downloads since April 2026; the methods it implements were endorsed at the 57th session of the UN Statistical Commission (March 2026).
  • R Shiny development — climate-health indicator modules delivered as Shiny applications with partner analysts, including one now operational at the National Institute of Statistics of Rwanda.
  • Technical lead for the AWS platform hosting both programmes: Docker and Kubernetes, CI in GitHub Actions, CD into a secure environment, semantic versioning, and monitoring for latency, error budgets and drift.
  • PySpark/Databricks pipelines for multi-source time-series and NLP workloads, with MLflow tracking, automated evaluation gates and data quality contracts.
  • Drafted the hosting contract that took the platform into production and saw the security assurance through to sign-off.
  • BREATHE (2025–2028) — own the analytical design and build of the ONS contribution to a Wellcome-funded programme led by the University of Bristol across eight institutions in five countries, running causal attribution over a 100-million-person cohort.
  • Line manage two data scientists and direct the technical work of statisticians, engineers and domain specialists across five countries. Set up the team’s GitHub organisation, onboarding pathway and engineering standards.

Independent AI Engineering

Part-time, alongside my main role — 2024 to present

  • RAG applications (LangChain, vector search) and predictive analytics (ARIMA, Prophet, LSTM), shipped as Dockerised services with automated retraining, evaluation harnesses and telemetry.
  • Prompt and version management, guardrails and human-in-the-loop review.
  • Most recent: the Power Planner, a seven-day electricity demand forecaster and supply planner, built live on stage at the NSE Glasgow 2026 conference.

Affiliate Researcher

James Watt School of Engineering, University of Glasgow — June 2023 to present

  • Federated anomaly detection and agentic AI on constrained hardware.
  • Engineering lead on a smart-home testbed run with a UK social landlord: a multi-sensor mould-risk forecasting system (24-hour AUC-ROC 0.851–0.957) and a privacy-preserving welfare monitoring system for lone elderly residents (F1 0.848/0.814, ~£200–300 per flat, no cameras, microphones or wearables).
  • Ongoing work across the Internet of Underwater Things, semantic telemetry, smart housing and ambient intelligence.

IoT Lead (2021–) & Data Science Lead (2024–)

Computer Science Academy Africa

  • Applied IoT, embedded systems and data science curricula delivered across 14 African countries; 200+ learners trained, teaching team of six.
  • Authored Python for IoT (2022) and Python for Data Science & Analysis (2024/2025).

Previous Roles

Data Scientist — Global Supply Chains Intelligence Programme

ONS Data Science Campus — January 2023 to May 2024

ONS technical focal point on the Analytical Working Group, embedded with customer departments across government.

  • Machine learning systems identifying companies using complex routing structures to breach Russian sanctions.
  • Analytical pipelines for customer departments: aerospace procurement for DSIT, bottleneck detection in critical medical reagent supply for the NHS, port-activity analysis of the impact of Brexit, and data quality remediation for DBT.
  • Predictive models of preferential tariff utilisation for UK imports, supporting trade negotiation — logistic regression, gradient boosting and network analysis over production networks and trade flows.
  • Transformer NLP for text classification, NER and buyer–supplier relation extraction; open-weight language models fine-tuned with QLoRA on GCP for industrial classification.
  • Distributed PySpark pipelines on Databricks and Cloudera with automated data quality monitoring; regular briefings to DBT, DSIT and Cabinet Office customers.

PhD Researcher

James Watt School of Engineering, University of Glasgow — 2019 to 2023

  • Cognitive networking for the Internet of Underwater Things, supervised by Prof Muhammad Ali Imran.
  • DEKCS, a clustering and routing protocol that prolonged underwater sensor-network lifetime by more than 70%.
  • Reinforcement-learning methods for simultaneous wireless information and power transfer (SWIPT) in underwater networks.

Editor, Engineering & Physical Sciences

Enago (2018–2020) and Cactus Communications (2017–2018) — several hundred manuscripts edited for international submission.