I build backends that stay up under load.

Scalable Python microservices on FastAPI, AsyncIO and gRPC inside an event-driven architecture — sustaining 850K daily requests at 99.9% uptime while cutting response latency 27%.

• Integrated ML and generative-AI endpoints for content recommendation, moderation and metadata summarization into backend services over REST, streamlining inference handling across creator-facing apps.
• Built caching and persistence layers with Redis, PostgreSQL and Cloud Storage for high-volume retrieval APIs — 33% less database load, 19% lower average query latency.
• Automated cloud-native delivery on GCP with Docker, Kubernetes and GitHub Actions, shortening release cycles 44% and hardening rollbacks.
• Rolled out distributed tracing with OpenTelemetry and Grafana Tempo, reducing mean time to diagnosis 36% during peak traffic.
• Hardened REST and LLM inference endpoints through PyTest suites, profiling and code review — 22% code-quality lift.

Modernized enterprise Java and Spring Boot services out of a legacy monolith into distributed, event-driven components — secure REST APIs processing 720K daily transactions, 31% faster.

• Wired scikit-learn fraud-detection models into Kafka-driven Python workflows, using feature engineering to cut false positives 24%.
• Tuned Redis caching alongside Oracle Database indexing and query optimization for high-traffic transaction endpoints — 26% lower latency, higher peak throughput.
• Automated OKE environments with Docker, Terraform and Jenkins pipelines: 28% better infrastructure utilization, 39% less release prep.
• Centralized security logging and audit trails in ELK and Splunk, flagging anomalous access and shortening investigations 37%.
• Raised API security standards with Spring Security and OAuth2 while practicing TDD with JUnit and Maven — 27% productivity gain.