https://mohakchugh.is-a.dev/blog Engineering notes by Mohak Chugh, SDE 2 at Amazon. en https://mohakchugh.is-a.dev/blog/iceberg-v3-deletion-vectors Tue, 07 Jul 2026 00:00:00 GMT Apache Iceberg v2 made row-level deletes possible with position delete files, and large deployments have regretted the details ever since. Format version 3 deprecates them in favor of deletion vectors, Roaring bitmaps stored in Puffin files with a hard invariant of at most one vector per data file. Here is what was broken, how the new binary format works down to the byte level, and why the length field is big-endian on purpose. https://mohakchugh.is-a.dev/blog/eagle-3-speculative-decoding-training-time-test Mon, 06 Jul 2026 00:00:00 GMT EAGLE-3 gets up to 6.5x decoding speedup by abandoning the feature-prediction objective that defined its predecessors. The interesting part is why feature prediction became the bottleneck, and how a trick called training-time test fixes the train/inference mismatch it leaves behind. https://mohakchugh.is-a.dev/blog/native-sparse-attention-hardware-aligned-long-context Mon, 06 Jul 2026 00:00:00 GMT DeepSeek's NSA makes attention sparsity a first-class citizen of pretraining instead of an inference-time hack, and pairs it with a kernel design that actually turns theoretical FLOP savings into wall-clock speedups. A close read of the architecture and why most sparse attention schemes before it failed to deliver. https://mohakchugh.is-a.dev/blog/rabitq-binary-quantization-vector-search Mon, 06 Jul 2026 00:00:00 GMT Product Quantization has powered billion-scale vector search for 15 years, but it can fail badly on real datasets and offers no theoretical guarantees. RaBitQ (SIGMOD 2024) compresses vectors to one bit per dimension, estimates distances with a popcount, and comes with a provable O(1/sqrt(D)) error bound. Here is how a random rotation makes that possible. https://mohakchugh.is-a.dev/blog/building-this-website Sun, 05 Jul 2026 00:00:00 GMT A technical walkthrough of rebuilding my portfolio from Angular 9 to Angular 22 with spartan-ng, Tailwind v4, SSG prerendering, a WOW animation system, and a zero-backend blogging engine powered by markdown files. https://mohakchugh.is-a.dev/blog/deterministic-simulation-testing Sun, 05 Jul 2026 00:00:00 GMT FoundationDB shipped a distributed database with essentially zero customer-reported bugs by running its entire cluster inside a single-threaded simulation driven by one random seed. TigerBeetle and Antithesis have since pushed the idea further. Here is how deterministic simulation testing works, why a seed is worth a thousand log files, and where the technique's real limits are. https://mohakchugh.is-a.dev/blog/prefill-decode-disaggregation-llm-serving Sun, 05 Jul 2026 00:00:00 GMT Prefill and decode have opposite hardware profiles, and serving them on the same GPUs wastes both. A practical tour of DistServe and Mooncake, the two papers behind the biggest architecture shift in LLM inference. https://mohakchugh.is-a.dev/blog/sieve-cache-eviction-simpler-than-lru Sun, 05 Jul 2026 00:00:00 GMT A 2024 NSDI paper showed that a FIFO queue, one bit per object, and a lazy hand pointer can out-perform LRU, ARC, and friends on web workloads, while removing the lock that makes LRU a scalability bottleneck. Here is how SIEVE works and why its simplicity is the whole point.