https://jatinbansal.com/Recent content in Jatin Bansal — Backend, Distributed Systems & AI Engineering on Jatin BansalHugoen-usSat, 16 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/query-transformations/Sat, 16 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/query-transformations/The query-side preprocessing layer for RAG: how rewriting, HyDE, multi-query, decomposition, and step-back prompting trade cost for recall.https://jatinbansal.com/ai-engineering/reranking/Thu, 14 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/reranking/Why cross-encoders dominate the precision stage of retrieval, when a reranker pays off, and how to compose cascades that respect the latency budget.https://jatinbansal.com/ai-engineering/hybrid-search/Wed, 13 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/hybrid-search/Why dense retrieval misses rare terms and exact matches, how BM25 and embeddings fuse via RRF, and the hybrid patterns that ship in production.https://jatinbansal.com/ai-engineering/chunking-strategies/Tue, 12 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/chunking-strategies/Why chunk size is RAG's most undertuned variable, how recursive, semantic, and structural chunking differ, and when parent-document retrieval wins.https://jatinbansal.com/ai-engineering/llm-inference-fundamentals/Mon, 11 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/llm-inference-fundamentals/How LLMs process text: BPE tokenization, the context window as working memory, KV caching, and sampling parameters that shape output variance.https://jatinbansal.com/ai-engineering/text-embeddings/Mon, 11 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/text-embeddings/How embedding models encode text as dense vectors, why cosine similarity captures meaning, and how to build semantic search in Python and TypeScript.https://jatinbansal.com/ai-engineering/vector-databases-ann/Mon, 11 May 2026 00:00:00 +0000https://jatinbansal.com/ai-engineering/vector-databases-ann/How HNSW, IVF, and ScaNN trade recall for speed, why exact KNN doesn't scale, and how to pick between pgvector, Qdrant, and Pinecone in production.https://jatinbansal.com/blog/event-loops-with-java-virtual-threads/Fri, 01 May 2026 00:00:00 +0000https://jatinbansal.com/blog/event-loops-with-java-virtual-threads/A practical guide to writing small event loops in Java 21 and Java 25 using virtual threads, blocking queues, direct control flow, and graceful shutdown.https://jatinbansal.com/blog/context-vs-prompt-engineering/Sun, 31 Aug 2025 00:00:00 +0000https://jatinbansal.com/blog/context-vs-prompt-engineering/Exploring the shift from prompt engineering to context engineering in AI systems, understanding context rot, and why managing context is becoming more critical than crafting prompts.https://jatinbansal.com/notes/claude-code-commands/Thu, 10 Jul 2025 00:00:00 +0000https://jatinbansal.com/notes/claude-code-commands/Comprehensive guide to Claude Code CLI commands including Docker MCP Gateway, authentication, scripting, advanced server management, and troubleshooting for efficient AI-powered development workflowshttps://jatinbansal.com/notes/deep-work/Sun, 06 Jul 2025 00:00:00 +0000https://jatinbansal.com/notes/deep-work/Deep work is a state of peak, distraction-free concentration that enables you to learn difficult things and produce high-quality work quickly.https://jatinbansal.com/notes/ai-llm-reading/Sat, 05 Jul 2025 00:00:00 +0000https://jatinbansal.com/notes/ai-llm-reading/A curated collection of resources that I find useful for learning about neural networks, LLMs, and AI development.https://jatinbansal.com/blog/stamped-lock/Sat, 05 Jul 2025 00:00:00 +0000https://jatinbansal.com/blog/stamped-lock/Learn how Java’s StampedLock enables near lock-free reads with optimistic locking, why it’s useful for virtual threads and read-heavy workloads, and how to use it safely.https://jatinbansal.com/blog/scaling-postgres/Tue, 05 Sep 2023 00:00:00 +0000https://jatinbansal.com/blog/scaling-postgres/Learn how we scaled PostgreSQL to handle millions of inserts per hour using application-level sharding with Spring Boot, combining table partitioning and host-level sharding for robust performance.