Essays and notes on enterprise AI strategy, governance, agentic platforms and safety-critical engineering.https://bbalaji.com/enhttps://bbalaji.com/writing/two-locks-on-a-static-site/https://bbalaji.com/writing/two-locks-on-a-static-site/This site is static, but it has a private studio where I write and keep documents. It is guarded twice: once at the edge by an identity service, and again inside the code by checking the signed token myself. Why one lock is not enough, and the ten tests that prove the second one.Sun, 22 Nov 2026 00:00:00 GMTSecurityArchitectureWebhttps://bbalaji.com/writing/zero-inline-scripts/https://bbalaji.com/writing/zero-inline-scripts/This site runs under a Content Security Policy that only allows scripts from its own files. Getting there took six fixes, most of them in places I did not expect: the build tool, a syntax highlighter, a page-turn library and a diagram renderer.Sun, 15 Nov 2026 00:00:00 GMTSecurityWebEngineeringhttps://bbalaji.com/writing/turning-html-into-a-book/https://bbalaji.com/writing/turning-html-into-a-book/My book is HTML. I wanted it to read like a book: real pages, a page turn, fullscreen that survives the next chapter, and a front page you can always get back to. How the reader measures pages in the browser, and what it costs.Sun, 08 Nov 2026 00:00:00 GMTWebBookEngineeringhttps://bbalaji.com/writing/an-ai-harness-that-cannot-lie/https://bbalaji.com/writing/an-ai-harness-that-cannot-lie/You cannot stop a language model from inventing a fact. You can build the harness around it so that an invented fact has nowhere to hide. The design of the small, private AI editor I use for documents that must stay true.Sun, 01 Nov 2026 00:00:00 GMTAgentic AIAI governanceArchitecturehttps://bbalaji.com/writing/editing-80000-words/https://bbalaji.com/writing/editing-80000-words/How I ran a full reader-first edit of a 17-chapter technical book with AI agents, without losing a code sample or a diagram. One brief, nine batches, a validator that compares every file against the original, and a human read at the end.Sun, 25 Oct 2026 00:00:00 GMTWritingBookAgentic AIhttps://bbalaji.com/writing/write-for-the-reader-not-the-auditor/https://bbalaji.com/writing/write-for-the-reader-not-the-auditor/Technical writing drifts into a defensive voice: passive, long, careful, addressed to nobody. I rewrote my 80,000-word book against nine rules and measured the result. Sentences got shorter, boxes got smaller, and the reader finally got addressed.Sun, 18 Oct 2026 00:00:00 GMTWritingBookTechnical communicationhttps://bbalaji.com/writing/why-the-title-says-mbse-for-ai/https://bbalaji.com/writing/why-the-title-says-mbse-for-ai/Months after naming my book, I could not remember why it read "MBSE for AI" and not the other way round. Working it out again gave me a simple test for naming any AI initiative: put the thing you are changing first.Sun, 11 Oct 2026 00:00:00 GMTSylangMBSEBookhttps://bbalaji.com/writing/ais-only-real-novelty-is-speed/https://bbalaji.com/writing/ais-only-real-novelty-is-speed/For most enterprises, AI does not do anything new. It does what we already knew how to do, at a cost and pace that change which work is worth doing. That is where strategy should start.Sun, 04 Oct 2026 00:00:00 GMTAI strategyEnterprise AIhttps://bbalaji.com/writing/tool-calling-is-the-benchmark-that-matters/https://bbalaji.com/writing/tool-calling-is-the-benchmark-that-matters/I benchmarked ten open models for agent work on my own GPU server and a laptop. Tokens per second got all the attention. The result that decided everything was whether the agent loop survived its second turn.Mon, 21 Sep 2026 00:00:00 GMTAgentic AILocal inferenceEvaluationhttps://bbalaji.com/writing/decks-as-data/https://bbalaji.com/writing/decks-as-data/Asking a model to produce a PowerPoint file gives you something hard to edit and impossible to govern. I built a deck generator that treats a presentation as structured data first. Sixteen components, twenty tests, and one rule that settled every argument.Mon, 14 Sep 2026 00:00:00 GMTGenerative AIArchitectureOpen sourcehttps://bbalaji.com/writing/find-the-value-stream-first/https://bbalaji.com/writing/find-the-value-stream-first/Most AI use cases are picked by asking people what they would like automated. I am building ValueAI to start somewhere else: map where the time actually goes, from four kinds of evidence, then decide where AI belongs.Mon, 31 Aug 2026 00:00:00 GMTAI strategyProcessValue stream mappinghttps://bbalaji.com/writing/rebuilding-sylang-on-the-data-platform/https://bbalaji.com/writing/rebuilding-sylang-on-the-data-platform/Sylang started as a text language in git, the way my old architecture language worked. This summer I moved its source of truth into the customer's own data platform. The decisions, the trade-offs, and the numbers behind them.Mon, 10 Aug 2026 00:00:00 GMTSylangMBSEArchitectureAgentic AIhttps://bbalaji.com/writing/isolation-is-not-a-prompt/https://bbalaji.com/writing/isolation-is-not-a-prompt/Running one agent framework for several users, I hit session and memory leaking between instances. The fix was not a better system prompt. It was operating-system isolation and a small control plane.Mon, 27 Jul 2026 00:00:00 GMTAgentic AISecurityArchitecturehttps://bbalaji.com/writing/a2a-is-a-contract/https://bbalaji.com/writing/a2a-is-a-contract/I registered two of my own agents in Gemini Enterprise over the A2A protocol. Registration failed on validation because the agent's responses were off-spec. What broke, how a small proxy fixed it, and what it means for agent interoperability in the enterprise.Mon, 20 Jul 2026 00:00:00 GMTA2AAgentic AIInteroperabilityhttps://bbalaji.com/writing/searching-requirements-without-sending-them-anywhere/https://bbalaji.com/writing/searching-requirements-without-sending-them-anywhere/Thousands of requirements and test cases, and a hard rule that none of it leaves the building. The self-hosted retrieval design I am testing, and the two choices that matter most.Mon, 13 Jul 2026 00:00:00 GMTRetrievalRequirementsData governancehttps://bbalaji.com/writing/small-models-dont-know-ears/https://bbalaji.com/writing/small-models-dont-know-ears/I tried open models up to 12B on a simple job: rewrite an existing requirement into EARS notation without inventing anything. None did it reliably. Here is the evaluation I built before training a 350M helper model.Mon, 06 Jul 2026 00:00:00 GMTRequirementsSmall modelsEvaluationhttps://bbalaji.com/writing/one-mcp-layer-instead-of-500-integrations/https://bbalaji.com/writing/one-mcp-layer-instead-of-500-integrations/When every AI client wants to talk to every engineering tool, the integration count grows as a product, not a sum. The arithmetic behind why I put one MCP layer in front of an enterprise engineering tool estate.Mon, 29 Jun 2026 00:00:00 GMTMCPAgentic AIArchitecturehttps://bbalaji.com/writing/why-i-wrote-a-book-on-mbse-for-ai/https://bbalaji.com/writing/why-i-wrote-a-book-on-mbse-for-ai/Every engineering company is asking which AI tool to bolt onto its modelling stack. I think that is the wrong question, and I wrote seventeen chapters on why. What is in the book, what is finished, and who it is for.Mon, 25 May 2026 00:00:00 GMTSylangMBSEBook