Led architecture and cross-team integration of backend systems supporting privacy-sensitive production iOS applications and public health programs, with emphasis on correctness, observability, low-latency performance, and operational reliability.

Led architecture and implementation of critical backend paths for LumiHealth’s financially sensitive transaction platform, defining server-side guarantees for an event-driven, append-only ledger system that prevented double spend, supported refunds and clawbacks, and preserved correctness and low-latency user experience under offline-client behavior and complex state transitions.

Defined shared engineering contracts across client, cloud, vendor, and government stakeholders through versioned specs, PlantUML diagrams, API contracts, and Protobuf-based interfaces, reducing ambiguity and enabling coordinated delivery under audited architecture constraints.

Led architecture for LumiHealth identity and backend integrations, including Singpass, translating Singapore government and GovTech privacy/security requirements into production controls for durability, correctness, and API performance in constrained network environments.

Led critical-path architecture and launch delivery for Apple’s Intuition study, implementing regulated HealthKit and SensorKit data flows from device to secure backend systems and helping enable launch on a compressed timeline.

Defined SLO-based operational practices across APAC and US coverage, including telemetry, alerting, incident response, and maintenance windows.

Built platform foundations that helped ML teams move from ad hoc experimentation to reproducible, deployable workflows across GCP and AWS environments.

Centralized ML infrastructure foundations across portfolio teams through infrastructure-as-code, deployment bootstrapping, and shared platform patterns.

Productionized ML workflows through dependency pinning, Docker packaging, and Kubernetes execution patterns.

Built standardized GPU-enabled experimentation environments with JupyterHub across GCP and AWS stacks.

Bridged ML experimentation to deployable infrastructure using MLflow and repeatable workflow patterns for startup engineering teams.

Led engineering for healthcare data and integration platforms in HIPAA-secured environments, spanning real-time ingestion, orchestration, and low-latency integration layers.

Built a Python-centric data platform supporting clinical and operational analytics.

Designed a real-time streaming ingestion and parsing system for live hospital EHR feeds powering clinical decision support.

Introduced and scaled Apache Airflow, standardizing DAG patterns, deployment practices, and on-call operations.

Built low-latency integration and ingestion layers, including Elasticsearch-based APIs and metadata-driven onboarding for 200+ heterogeneous EHR sources.