Working on a payment gateway platform with live products in Costa Rica and Kenya. Responsible for backend reliability, performance engineering, and developer tooling.

• Optimized main list APIs by eliminating N+1 query patterns — reduced response time from ~4s to 0.5s.
• Designed and built a Kafka-based data pipeline to migrate data from Couchbase to MySQL, reducing processing time from ~8 hours to ~5 minutes.
• Refactored legacy services: extracted non-critical logic from transactions and applied retry logic with exponential backoff.
• Designed new CMS API set that consolidated 21 endpoints to 9, simplifying maintenance and developer experience.
• Enhanced Kafka consumers with retry logic, structured logging, and failed-message tracking for manual reprocessing.
• Upgraded Python base image from 3.7 (EOL) to 3.12 across all services.
• Set up Jenkins CI pipeline; raised test coverage to 80%+ on brownfield services and 90%+ on greenfield services, reducing reported bugs by 70%.

Joined as individual contributor on a data platform for epilepsy research; promoted to tech lead managing a 5-person team. Partnered directly with Australian data scientists and clinicians to translate requirements into technical solutions.

• Built REST APIs for patient, study, and report management using AWS API Gateway + Lambda + PostgreSQL.
• Designed and implemented a real-time WebSocket notification system using AWS API Gateway WebSocket APIs, Lambda, and SQS.
• Built a PDF generation pipeline: ETL from REDCap via EventBridge + Lambda, rendering with PyMuPDF, storage on S3.
• Deployed multiple ML models (regression, XGBoost, NLP) to AWS SageMaker real-time endpoints.
• Promoted to tech lead: led planning, code reviews, mentoring, and knowledge transfer for a 5-person team.
• Produced proof-of-concept demos directly with Australian data scientists and clinicians.