How I Replaced My $20/Month Claude Subscription With Gemma 4 Running Locally Through OllamaUsing ollama launch claude with Gemma 4 for zero-cost agentic AI coding — full setup, capabilities deep-dive, and honest review11 min read·shivamshasheesh.hashnode.devApr 4
Building Semantic Search Over Automotive AUTOSAR Configurations — How I Made 1000+ XML Parameters Queryable in Natural LanguageDesigning an ETL + RAG pipeline to parse complex ARXML schemas, generate 3072-dim embeddings, and answer engineering questions about ECU configs instantly6 min read·shivamshasheesh.hashnode.devFeb 10
Google Stitch + MCP Server Changed How I Design UI — Here's My Full WorkflowUsing Stitch's AI-native canvas and MCP server to go from natural language to production-ready UI designs in minutes9 min read·shivamshasheesh.hashnode.devDec 15, 2025
I Built a Production-Grade RAG Framework With Hybrid Search, Contextual Chunking, and Zero Cloud Lock-InHow ragpipe combines BM25 + dense vectors, Anthropic's contextual retrieval, HyDE query expansion, and Ollama-first local inference into a modular Python framework9 min read·shivamshasheesh.hashnode.devOct 20, 2025
