Practical engineering notes on distributed systems, databases, data infrastructure and AI agents.https://kharekartik.dev/en-ushttps://kharekartik.dev/writing/my-llm-query-optimizer-demoed-in-a-week-making-it-honest-took-months/https://kharekartik.dev/writing/my-llm-query-optimizer-demoed-in-a-week-making-it-honest-took-months/Why a week-long LLM query optimizer demo needed months of evals, validators and live cluster probes before its advice became trustworthy.Fri, 10 Jul 2026 00:00:00 GMThttps://kharekartik.dev/writing/the-staff-engineers-missing-manual/https://kharekartik.dev/writing/the-staff-engineers-missing-manual/A practical field guide to the staff engineer transition: turning ambiguity into durable progress without trying to solve every problem yourself.Wed, 08 Jul 2026 00:00:00 GMThttps://kharekartik.dev/writing/i-am-building-a-qa-agent-because-coding-got-too-fast/https://kharekartik.dev/writing/i-am-building-a-qa-agent-because-coding-got-too-fast/An in progress dump from trying to turn AI coding speed into runtime proof instead of a bigger pile of plausible green checks.Mon, 25 May 2026 00:00:00 GMThttps://kharekartik.dev/writing/me-vs-rocksdb-in-production/https://kharekartik.dev/writing/me-vs-rocksdb-in-production/What started breaking once RocksDB crossed from useful embedded store into a scale-sensitive production subsystem.Wed, 20 May 2026 00:00:00 GMThttps://kharekartik.dev/writing/wtf-does-it-take-to-automate-visual-explainers-part-2/https://kharekartik.dev/writing/wtf-does-it-take-to-automate-visual-explainers-part-2/The harness could finally build something, but the first real simulators taught me that better prompts do not fix a visual system with too much freedom.Sat, 16 May 2026 00:00:00 GMThttps://kharekartik.dev/writing/how-im-using-claude-code-to-reduce-my-workload/https://kharekartik.dev/writing/how-im-using-claude-code-to-reduce-my-workload/A growing set of small Claude Code automations, each tied to a specific chore I already do. A living list I'll keep adding to as I build more.Tue, 21 Apr 2026 00:00:00 GMThttps://kharekartik.dev/writing/building-production-ai-agents-incrementally/https://kharekartik.dev/writing/building-production-ai-agents-incrementally/What I learned building a Jira-to-PR agent for Apache Pinot one proven layer at a time instead of designing the whole system upfront.Thu, 16 Apr 2026 00:00:00 GMThttps://kharekartik.dev/writing/wtf-does-it-take-to-automate-visual-explainers-part-1/https://kharekartik.dev/writing/wtf-does-it-take-to-automate-visual-explainers-part-1/I spent a year hand-building visual explainers in Cursor. Then I tried to automate it with AI. The first month was all agent infrastructure and zero visuals.Sat, 11 Apr 2026 00:00:00 GMThttps://kharekartik.dev/writing/debugging-race-conditions-in-distributed-systems/https://kharekartik.dev/writing/debugging-race-conditions-in-distributed-systems/A real incident walkthrough for when your system is healthy and completely wrongThu, 02 Apr 2026 00:00:00 GMThttps://kharekartik.dev/writing/wtf-is-time-travel-in-data-lakes-and-does-it-actually-solve-anything/https://kharekartik.dev/writing/wtf-is-time-travel-in-data-lakes-and-does-it-actually-solve-anything/A practical explanation of snapshots, Delta Lake, Apache Iceberg and why data teams suddenly care about table formats.Tue, 31 Mar 2026 00:00:00 GMThttps://kharekartik.dev/writing/what-arrow-actually-demands-from-you/https://kharekartik.dev/writing/what-arrow-actually-demands-from-you/The art of removing costs you didn't know you were paying on every row.Tue, 12 Aug 2025 00:00:00 GMThttps://kharekartik.dev/writing/anime-pfp-chrome-extension/https://kharekartik.dev/writing/anime-pfp-chrome-extension/What it took to make a Stable Diffusion pipeline fast enough to replace Twitter profile pictures with anime versions as you scroll.Thu, 20 Jun 2024 00:00:00 GMThttps://kharekartik.dev/writing/apache-helix-the-distributed-system-s-orchestra-conductor/https://kharekartik.dev/writing/apache-helix-the-distributed-system-s-orchestra-conductor/Achieve harmony in complex clusters using finite-state machinesTue, 28 Feb 2023 00:00:00 GMThttps://kharekartik.dev/writing/navigating-the-minefield-of-rocksdb-configuration-options/https://kharekartik.dev/writing/navigating-the-minefield-of-rocksdb-configuration-options/Unleashing the full potential of your RocksDB with the right configurationTue, 03 Jan 2023 00:00:00 GMThttps://kharekartik.dev/writing/how-to-package-java-projects-in-python-tar-files/https://kharekartik.dev/writing/how-to-package-java-projects-in-python-tar-files/How to package a Java project inside a Python distribution when the two language ecosystems genuinely need to ship together.Tue, 02 Mar 2021 00:00:00 GMThttps://kharekartik.dev/writing/utilize-udfs-to-supercharge-queries-in-apache-pinot/https://kharekartik.dev/writing/utilize-udfs-to-supercharge-queries-in-apache-pinot/How Apache Pinot's Groovy UDF support extends SQL queries with custom functions and the tradeoffs that come with it.Tue, 29 Sep 2020 00:00:00 GMThttps://kharekartik.dev/writing/leverage-plugins-to-ingest-parquet-files-from-s3-in-pinot/https://kharekartik.dev/writing/leverage-plugins-to-ingest-parquet-files-from-s3-in-pinot/How Apache Pinot's plugin architecture lets ingestion jobs read Parquet data from Amazon S3 through pluggable filesystems and input formats.Tue, 18 Aug 2020 00:00:00 GMThttps://kharekartik.dev/writing/learning-multi-dimensional-indices-the-next-big-thing-in-olap-dbs/https://kharekartik.dev/writing/learning-multi-dimensional-indices-the-next-big-thing-in-olap-dbs/How multi-dimensional indexes accelerate OLAP queries by organizing data around the dimensions filters actually use.Thu, 09 Apr 2020 00:00:00 GMThttps://kharekartik.dev/writing/a-glimpse-into-my-wfh-in-quarantine-life/https://kharekartik.dev/writing/a-glimpse-into-my-wfh-in-quarantine-life/A practical tour of the tools, routines and home-office setup that shaped my work-from-home life during quarantine.Wed, 01 Apr 2020 00:00:00 GMThttps://kharekartik.dev/writing/how-does-zookeeper-servers-remain-in-sync/https://kharekartik.dev/writing/how-does-zookeeper-servers-remain-in-sync/How ZooKeeper keeps leader and follower servers synchronized through atomic broadcast, transaction logs and snapshots.Mon, 30 Mar 2020 00:00:00 GMThttps://kharekartik.dev/writing/why-apache-airflow-is-a-great-choice-for-managing-data-pipelines/https://kharekartik.dev/writing/why-apache-airflow-is-a-great-choice-for-managing-data-pipelines/A glimpse at capabilities which makes Airflow better than its predecessorsMon, 20 Jan 2020 00:00:00 GMThttps://kharekartik.dev/writing/deploying-ml-models-in-distributed-real-time-data-streaming-applications/https://kharekartik.dev/writing/deploying-ml-models-in-distributed-real-time-data-streaming-applications/Explore the various strategies to deploy ML models in Apache Flink/Spark or other realtime data streaming applications.Sat, 11 Jan 2020 00:00:00 GMT