Po Jui Lin

About

Backend engineer with experience at Amazon building an API-driven service (APS) for large-scale prompt optimization jobs using AWS Step Functions, Batch/Fargate, DynamoDB, and S3, emphasizing idempotency, observability, and secure execution boundaries. Also led a multi-tenant enterprise policy/configuration backend refactor at MAMIT Cyber with versioned schemas, shadow writes, feature-flagged rollout, and PostgreSQL RLS-based tenant isolation.

Key Strengths

• Designed large-scale job orchestration backend on AWS (Step Functions + Batch/Fargate + DynamoDB + S3) with focus on scalability, reliability, and cost efficiency
• Strong idempotency and state-management patterns (run IDs, conditional writes, deterministic outputs) to make retries safe and prevent duplicate/partial results
• End-to-end observability implementation across distributed components (structured logs, correlation IDs, artifact traceability)
• Security-first backend design (least-privilege IAM, role separation, tenant isolation, PostgreSQL RLS with default-deny policies)
• API contract-first mindset that improves frontend enablement (stable models, UI-friendly state machines, fewer round trips, explicit error codes)
• Led controlled backend migrations with risk mitigation (versioned schemas, compatibility layers, shadow writes, feature-flagged tiered rollout, validation for semantic equivalence)
• Proactively identified and fixed subtle failure modes (cancellation/zombie runs, finalization validation) to prevent corrupt outputs

Experience

Education

Awards

• Dean’s List (2017) - Hong Kong University of Science and Technology

Skills

Languages