I transform complex architectural challenges into predictable, ultra-low-latency distributed services. From microsecond-level concurrency to event-driven streaming, I engineer production backends designed to scale effortlessly to millions of requests.
Interactive real-time topology of my production Go backend — simulating high-throughput WebRTC video/audio streaming with LiveKit SFU, event ingestion via Apache Kafka, and sub-millisecond Redis Pub/Sub broadcast.
Navigate to dedicated in-depth sections covering live web platforms, engineering philosophy, distributed tech stack internals, and production solutions.
Discover the 4 live standalone web applications in the seang.org cloud — FitViet Web, Doc Tool, Dev Tool, and Lα»ch Γm — built for sub-second performance and zero-server privacy.
Explore Ecosystem →Mechanical sympathy, the concurrency mindset, zero data races, and strict memory budgeting. The core architectural values guiding how I engineer resilient distributed software.
Read Philosophy →Production-hardened technologies: Golang runtime internals (GMP, sync.Pool, GC), Apache Kafka, Redis clustering, LiveKit WebRTC SFU, Kubernetes, and PostgreSQL tuning.
Inspect Tech Stack →Six mission-critical case studies: 500k+ RPS Go APIs, low-latency LiveKit streaming, Kafka EDA pipelines, PgBouncer sharding, and memory-bounded zero-OOM file streaming.
View Case Studies →Production-tested distributed systems, Golang concurrency internals, low-latency consensus, and resilient cloud architecture engineered by Marcus.
A deep-dive into bounded worker pool architecture in Go to process millions of concurrent tasks per second without unbounded goroutine memory leaks.
Analyzing critical failure modes in distributed locking: why naive SETNX breaks under network partitions, Martin Kleppmann's critique, and fencing tokens.
Mastering three classic cache failure modes: cache breakdown/stampede prevention with Go singleflight, Bloom filters, and jittered TTL expiration.