I build large-scale distributed systems, AI infrastructure, and GPU compute platforms — from streaming fault detection with Flink and Kafka to cloud services operating at million-QPS scale.
Principal engineer focused on distributed systems, databases, cloud infrastructure, GenAI, and AI infrastructure. I own systems from architecture through production, balancing scalability, reliability, security, performance, and cost.
AI Infrastructure · GPU Observability · Developer Tooling
Led Big Data Gateway architecture and production launch; designed capacity for ~1M QPS while reducing compute footprint from 85 to 32 nodes per region.
Built multi-tenancy, infrastructure automation, and cross-account VPC connectivity for distributed database workloads.
Built production backend systems and microservices for high-traffic consumer and seller platforms.
Object Storage → Apache Flink → fault detection → Apache Kafka → downstream customer workflows.
Fault-tolerant agent execution with idempotent retries, event-sourced durability, deterministic replay, concurrency control, and tracing.
PagedAttention-style KV cache allocator with copy-on-write, prefix sharing, eviction policies, and benchmarking.