--- title: "Pragnesh Shah (@prag-matic-builder) | Hashnode" description: "Bridging Enterprise Legacy and GenAI-Native Futures | Speaker and Strategist. Read the latest articles by Pragnesh Shah on Hashnode." url: "https://hashnode.com/@prag-matic-builder" --- @prag-matic-builder # Pragnesh Shah @prag-matic-builder · London, UK · Joined May 2026 Bridging Enterprise Legacy and GenAI-Native Futures | Speaker and Strategist ### [Query Locally, Think Remotely: How DuckDB-Wasm Cuts LLM Token Cost](https://pragmatic-builder.hashnode.dev/query-locally-think-remotely-cuts-llm-token-cost?utm_source=hashnode&utm_medium=feed) Learn how DuckDB-Wasm queries CSV files in the browser, keeps raw data local, and sends compact SQL results to AI agents to cut LLM token cost 17 min read · [pragmatic-builder.hashnode.dev](https://pragmatic-builder.hashnode.dev/query-locally-think-remotely-cuts-llm-token-cost?utm_source=hashnode&utm_medium=feed) [Aug 4](https://hashnode.com/posts/query-locally-think-remotely-cuts-llm-token-cost/6a91b2cbac000cd52d33fa10) ### [I Used DeepSeek R1 + Open Data to Challenge AI ROI](https://pragmatic-builder.hashnode.dev/deepseek-r1-open-data-ai-roi?utm_source=hashnode&utm_medium=feed) I used DeepSeek R1 on AWS Bedrock to test AI ROI claims against SEC revenue data and developer productivity 11 min read · [pragmatic-builder.hashnode.dev](https://pragmatic-builder.hashnode.dev/deepseek-r1-open-data-ai-roi?utm_source=hashnode&utm_medium=feed) [Jun 4](https://hashnode.com/posts/deepseek-r1-open-data-ai-roi/6a21827f4812d84ddc9eed8d) ### [I Built a RAG System That Actually Works — Hybrid Search, Reranking, and the Decisions That Mattered](https://pragmatic-builder.hashnode.dev/rag-hybrid-search-reranking-decisions?utm_source=hashnode&utm_medium=feed) How I built an advanced RAG pipeline with hybrid search, reranking, and semantic chunking — the architectural decisions and code that made it work. 8 min read · [pragmatic-builder.hashnode.dev](https://pragmatic-builder.hashnode.dev/rag-hybrid-search-reranking-decisions?utm_source=hashnode&utm_medium=feed) [May 31](https://hashnode.com/posts/rag-hybrid-search-reranking-decisions/6a1c97ff072f333f8c51be1a)