Luke's BlogPersonal blog of Luke Shenhttps://ulixert.github.io/Building Theseon: Architecture of a Distributed LSM and Vector Engine in Gohttps://ulixert.github.io/posts/building-theseon/https://ulixert.github.io/posts/building-theseon/Architecture and design decisions behind a hand-built distributed LSM engine with HNSW vector search: from skip lists and SSTables to SWIM gossip, quorum coordination, and approximate nearest neighbors.Thu, 08 Jan 2026 00:00:00 GMTSnapshots, Transactions, and the Art of Not Blocking Writershttps://ulixert.github.io/posts/theseon-mvcc-transactions/https://ulixert.github.io/posts/theseon-mvcc-transactions/How Theseon gained MVCC snapshot isolation and optimistic transactions — by separating merge from dedup and making versioning a first-class feature.Sun, 22 Mar 2026 00:00:00 GMTQuorum Reads, Quorum Writes, and the Repair That Followshttps://ulixert.github.io/posts/theseon-quorum-coordinator/https://ulixert.github.io/posts/theseon-quorum-coordinator/How Theseon's coordinator fans out operations to replicas, resolves conflicts by HLC timestamp, and repairs stale data in the background — plus the latency traps and test design bugs that emerged.Tue, 31 Mar 2026 00:00:00 GMTBuilding HNSW from Scratch: Graph Construction, Beam Search, and What Recall Actually Measureshttps://ulixert.github.io/posts/theseon-hnsw-scratch/https://ulixert.github.io/posts/theseon-hnsw-scratch/Implementing the HNSW approximate nearest neighbor algorithm from the original paper in Go — the insert and search algorithms, neighbor selection heuristics, tombstone-aware traversal, and building an evaluation framework to prove it works.Sat, 04 Apr 2026 00:00:00 GMTThe Storage Foundation: Memtable, WAL, and SSTableshttps://ulixert.github.io/posts/theseon-storage-foundation/https://ulixert.github.io/posts/theseon-storage-foundation/A deep dive into the bottom half of the Theseon stack: data structures and on-disk formats.Thu, 12 Mar 2026 00:00:00 GMTBuffering Writes for Dead Replicas: Hinted Handoffhttps://ulixert.github.io/posts/theseon-hinted-handoff/https://ulixert.github.io/posts/theseon-hinted-handoff/How Theseon stores writes destined for dead nodes and replays them on recovery — the capacity accounting race, the iterator deadlock, and why skipping fsync is the right call for ephemeral data.Thu, 09 Apr 2026 00:00:00 GMTSequence Numbers, the Merge Iterator, and Wiring It All Togetherhttps://ulixert.github.io/posts/theseon-wiring-it-together/https://ulixert.github.io/posts/theseon-wiring-it-together/How the storage foundation pieces are wired into a working engine, including internal key encoding and the merge iterator.Sun, 15 Mar 2026 00:00:00 GMTWho's Alive? Building SWIM Failure Detection from Scratchhttps://ulixert.github.io/posts/theseon-swim-protocol/https://ulixert.github.io/posts/theseon-swim-protocol/How Theseon detects node failures without a leader — implementing the SWIM gossip protocol from the paper, the bugs that emerged, and why liveness and data ownership must be decoupled.Sun, 29 Mar 2026 00:00:00 GMTMaking Vectors Durable: KV Integration, Snapshot Persistence, and the Bugs Along the Wayhttps://ulixert.github.io/posts/theseon-vector-kv-integration/https://ulixert.github.io/posts/theseon-vector-kv-integration/Storing HNSW vectors as regular LSM entries in Theseon — the encoding format, key layout, per-collection concurrency model, a graph connectivity bug, and binary snapshot persistence to avoid full rebuilds on restart.Tue, 07 Apr 2026 00:00:00 GMTMaking the Engine Self-Maintaining: Compaction, Caching, and Durabilityhttps://ulixert.github.io/posts/theseon-self-maintaining/https://ulixert.github.io/posts/theseon-self-maintaining/How the manifest, leveled compaction, block cache, and write batches turned Theseon into a self-maintaining storage engine.Wed, 18 Mar 2026 00:00:00 GMT