Responsibilities & Impact
- Led backend development with Java & MongoDB for a healthcare prototype
within an Agile environment, enabling
efficient analysis of patient lab reports as part of a client project through the Centre for
Digital Media.
- Developed Python scripts for rapid prototyping to preprocess OCR output and
test different LLMs for patient data
extraction from lab reports, version-controlled with GIT to ensure reproducibility and
smooth collaboration, accelerating
evaluation and integration with Azure OpenAI.
- Reduced setup time and enabled easy cloud deployment for the Providence HC
project by deploying and containerizing
backend services with Docker and Fly.io.
- Reduced user workflow steps by 67% by creating an automated
process that replaced multiple manual actions, resulting in a significantly improved user
experience and effectively showcasing the prototype’s potential.
- Integrated Generative AI capabilities using Azure
OpenAI to address the challenge of interpreting test reports, collaborating
with team members to refine LLM prompts, resulting in better extraction accuracy of patient
data.