https://ndthuan.com/{ndt.write();}Software engineer exploring AI, software systems, and how technology changes the way we work. 2026-10-03T01:28:17+07:00 Thuan Nguyen https://ndthuan.com/ Jekyll © 2026 Thuan Nguyen /assets/img/favicons/favicon.ico /assets/img/favicons/favicon-96x96.png The GIL Bottleneck Hiding in Self-Hosted LLM Serving2026-09-25T12:00:00+07:00 2026-09-25T12:00:00+07:00 https://ndthuan.com/gil-bottleneck-self-hosted-llm-serving/ Thuan Nguyen Why a custom logits processor that runs once per request can dominate tail latency even when the GPU batches perfectly, and the batch-level rewrite that fixes it. Open-weight LLMs are winning on cost, not benchmarks2026-09-13T10:00:00+07:00 2026-09-13T10:00:00+07:00 https://ndthuan.com/open-weight-models-winning-on-cost/ Thuan Nguyen Open-weight models are not catching up to frontier APIs on benchmarks. They are already ahead on the number that decides most production budgets: cost per request at scale. The Transactional Outbox Pattern: Reliable Events Without Distributed Transactions2026-09-04T09:00:00+07:00 2026-09-04T09:00:00+07:00 https://ndthuan.com/transactional-outbox-pattern/ Thuan Nguyen Saving to the database and publishing an event are two writes to two systems that fail independently. The transactional outbox pattern makes the event part of the same database transaction, so what commits is what gets published - here is a working Go and PostgreSQL implementation, including the retry, ordering, latency and cleanup details most guides skip. A Weekend Ride to Vũng Tàu, and Eight Walls I Came Home With2026-08-27T00:00:00+07:00 2026-08-27T00:00:00+07:00 https://ndthuan.com/vung-tau-by-bicycle/ Thuan Nguyen A weekend bicycle ride to Vũng Tàu, the road that made it worth doing, and eight photographs of walls. Notes on riding for the pleasure of it, and on why a town you arrive in slowly gives you different pictures. Bloom Filters: How Databases Skip Disk Reads They Don't Need2026-08-24T00:00:00+07:00 2026-08-24T00:00:00+07:00 https://ndthuan.com/bloom-filters-how-databases-skip-disk-reads/ Thuan Nguyen Bloom filters answer one narrow question - definitely not present, or maybe present - in a fixed amount of memory, and that narrow answer is enough to save databases from millions of disk reads for keys that were never there. Here is how they work, the math behind the false positive rate, and where RocksDB, Cassandra, and PostgreSQL actually use them.