Senior Software Engineer

• With an 8-strong team we put together the principal clinical trial platform in Angular, making use of standalone components and signals. A design system was put in place from the ground up; by way of Tailwind we have styled a combination of Angular Material and our own Radix UI components to produce 50 or so that are to be found on adverse event reports, consent forms and the like on the trial dashboards. • For the patient portal we turned to Next.js and its App Router so as to server-side render educational content and consent forms, a move that has made for better performance and accessibility in a healthcare setting. The shadcn/ui and Tailwind CSS give us a reusable design system in keeping with current UI standards. We see a 60% gain in initial page load times thanks to the way we have optimized the application: static generation, server-side rendering and React Query for optimistic updates and managing server state. RxJS is employed for any complex data streams while NgRx provides centralised control over state in enterprise clinical workflows. • ASP.NET Core microservices serve as the backend for our clinical data processing, with C# and Entity Framework Core handling the domain logic, while Java Spring Boot services manage protocol execution and Node.js/Express handles what is done in real time. In all we put through 200,000 clinical data points a day in services dedicated to protocol execution, adverse events and patient management. SignalR and Socket.IO have been added for team messaging and remote monitoring. • The SQL Server and PostgreSQL schemas were designed with the right relationships and indexing to cope with the volume of trial data. To offload the database by 45 per cent we have Redis running message queues with MassTransit (.NET) and Celery to take care of background jobs. Data at rest is encrypted in the pipeline so as to remain HIPAA-compliant. • AWS and Azure are where we have deployed, with EC2 and Azure App Service for the back end, S3 and Azure Blob Storage to store documents and Lambda/Azure Functions for any workflow automation. SQS/SNS and Azure Service Bus are configured for the async side of adverse event notifications and CloudFront provides global CDN. Our infrastructure is set up to the principle of least privilege in accordance with HIPAA. • On EKS and AKS we orchestrate the microservices, complete with namespace isolation and horizontal pod autoscaling. GitHub Actions and Azure DevOps form the CI/CD pipeline and will run xUnit, Pytest, Jest and Cypress without intervention. Deployments are made consistent in every environment via Helm charts and Docker multi-stage builds. • Devised intelligent automation to classify adverse events and spot protocol deviations, making use of OpenAI GPT APIs in conjunction with LangChain. Development was put through its paces with the aid of Cursor IDE and Claude Code. A Pinecone vector database was also put in place for semantic search purposes, which has seen a 60 per cent gain in the efficiency of data quality reviews. • FHIR APIs have been put in to interface with hospital EHRs and for the parsing of HL7 from legacy systems. Electronic signatures, role-based access and full audit trails ensure conformance to GxP, GAMP4 and HIPAA, with validated environments as per 21 CFR Part 11.