Professional Projects

A selection of work across machine learning, data quality, and large-scale backend services. Roles at Netcompany delivering for the NHS.

Relationship Network (NHS PDS)

May 2024 – Present — Leading development of a relationship network datastore for NHS PDS and exposing relationships via a RESTful FHIR API to enable proxy care access in the NHS App for 4.5M+ children (plus 350K new children per year).

Stack: AWS (Lambda, Step Functions, SQS, S3, API Gateway, DynamoDB, Neptune), Python (Pydantic, OpenAPI), Terraform, Swagger/OpenAPI
Highlights: Designed ETL ingestion from GRO birth registrations into Neptune; implemented a leaky bucket rate limiter for PDS FHIR API; building the RelatedPerson FHIR API
Role: Senior Software Engineer — design, implementation, stakeholder alignment, agile ceremonies

Address Quality Service (Autofix)

Oct 2023 – May 2024 — Created an API in AWS to enrich NHS PDS address data with UPRNs by integrating with OS Places and NHS Spine II. Deployed via IaC and integrated into core Spine workflows safely.

Stack: AWS (API Gateway, Glue, Lambda, SQS), Python, Terraform; Integration with NHS Spine II
Impact: Added 1.5M UPRNs and continues to apply 50,000+/month; saving the NHS £200,000+/year in avoided letters
Testing: Integration and performance tests to ensure zero disruption to existing workflows

Duplicate Patient Classification (DQIP)

Jan 2023 – Oct 2023 — Led a data science initiative to automate duplicate record triage for back-office staff, delivering a classification model that saves 60+ seconds per case and 4,000+ hours/year.

Stack: Databricks (SQL, PySpark UDFs), Python (NumPy, Pandas, SciPy, scikit-learn)
Work: Data extraction and cleansing, feature engineering, modeling and evaluation; workshops to capture assumptions
Impact: Estimated savings of £80,000+/year; improved data integrity and reduced backlog

NHS Spine II — Feature Development

Dec 2021 – Dec 2022 — Full‑stack development on a mission‑critical legacy system. Shipped features with end‑to‑end tests and close collaboration with platform teams.

Stack: Python (Flask), Jinja2, HTML/CSS/JS (jQuery), Linux; Jenkins, Splunk; tests with PyTest, Selenium; E2E via XML
Practice: Agile ceremonies, code reviews, integration/performance testing