Raghunandan Gupta·9h agoWhat the AI Assistant Demo Hid: Five Days From Demo to ShipAsk Your Portal · Post 3 of 18: the failures real users found first, and the limits we still ship with.
Raghunandan Gupta·9h agoThe Twenty-Line Agent Loop Behind a Read-Only AI AssistantAsk Your Portal · Post 2 of 18: the loop, one tool and the read-only helper, as code you can copy.
Raghunandan Gupta·1d agoOne Tool Registry, Many DoorsAsk Your Portal · Post 4 of 18: Why the tool registry, not the chat page, is the product, and what one tool needs
Raghunandan Gupta·1d agoBuild an In-App AI Assistant in HoursAsk Your Portal · Post 1 of 18: the three decisions to make before any code, and an architecture you can reuse.
Raghunandan Gupta·Oct 21Designing Multi-Agent Frameworks: Why This Series ExistsSeries introduction for engineers building production-grade agentic systems
Raghunandan Gupta·Oct 18Inside the Ingestion Engine: Turning Messy A/B Experiment Reports Into Processable DataPart 2: Exploring a document-processing architecture with content hashing, asynchronous orchestration, multi-format parsing, and…A response icon1A response icon1
Raghunandan Gupta·Oct 18The A/B Testing Blind Spot: Building a Knowledge Base Beyond Naive RAGTurn scattered A/B test docs into structured, reusable team memory with an AI-powered knowledge base built for continuous learning.dsds
Raghunandan Gupta·Oct 17Series: Building an AI-Powered Knowledge Base for ExperimentationFrom fragmented experiment documents to structured knowledge, semantic search, and AI-assisted decision-making.
Raghunandan Gupta·Jun 21Building a Resilient Async AI Pipeline: Optimizations That Cut Our Failure Rate by 50%We run an async pipeline that processes uploaded documents through multiple AI steps: image analysis, content extraction, multimodal…
Raghunandan Gupta·Jun 14Every Design Doc I Reviewed Last Year Had the Same 5 Gaps.Download Review Genie Chrome extension