I'm passionate about creating efficient software systems and fostering positive team dynamics. I enjoy viewing software engineering problems with a creative lens, blending design and engineering principles to solve them.
Spearheaded the Dataset Management project, resulting in improved ML Engineer workflows, enhanced model quality, and a 65% reduction in duplicated assets.
•
Developed a scalable Celery application, expanding routing capacity from <1,000 to 50,000+ multi-gigabyte slide files/month, enabling PHI-sensitive file de-identification and traceability.
•
Refined user stories that boosted sprint accuracy and increased successful completion of story points from 50% to 80%.
•
Designed and implemented a process to identify and remove 1.5+ million instances of sensitive data across PostgreSQL, AWS S3, and AWS EFS, while reducing estimated processing time from over 3 weeks to just 3 days utilizing Kubernetes StatefulSets.
•
Designed an automatic semantic-versioning based release process for an FDA-validated service to ensure compliance, significantly reducing manual engineering validation time from 2 days to under 5 hours.
•
Mentored and onboarded nine engineers, optimized the onboarding process, and fostered career growth for junior engineers. Actively contributed to recruitment initiatives and successfully guided several promotions.
Designed a Flask-based ML model metric-aggregation micro-service and introduced OpenAPI code generation, streamlining development and significantly reducing the time required for ML Engineer model evaluation metric gathering and report generation from 8 hours to near-instantaneous.
•
Authored design documentation that reduced slide-data processing runtimes by 78%, cutting DZI tile storage requirements by 48% and significantly lowering AWS S3 costs.
•
Led a large refactor of a legacy application. Refactored the Flask backend to Onion architecture and modernized the Vue frontend by leading the push to TypeScript and Vue Composition API.
•
Fostered a collaborative team culture through regular code pairing, learning sessions, and engineering practice meetings. Established an operations and release rotation, enhancing coverage for end-user issues.
•
Championed tech-debt reduction by managing a technical debt epic, achieving a 61% decrease in backlog bugs.
Designed and executed an email-marketing project which generated a new revenue stream and affected 350,000 users daily. Collaborated with five stakeholder teams and presented final results at an engineering-wide meeting.
•
Added capability for various data retrieval variants to be used in PMML recommendation models, including a vector solution which sped up database time from 100+ms to less than 10ms.
•
Coordinated Tech Talks for the Hotels business unit, engaging over 60 regular attendees.
•
Enhanced price outlier detection using geographic standard-deviation filtering in a PySpark ETL, increasing the accuracy of max-price data per location and generating revenue by attracting long-tail customers.
•
Led pre-requisite planning and executed the necessary changes for an older micro-service to be ready for CI/CD, which changed the deployment cycle from 2 weeks to any moment.
•
Developed in a full-stack environment using Java and ReactJS.
•
Built several location-based ETLs responsible for migrating millions of rows of data daily.
