Notes on training engineers, forward-deployed work and putting AI into real workflows.https://mayurpatil.ai/https://mayurpatil.ai/writing/why-forward-deployed-engineering/https://mayurpatil.ai/writing/why-forward-deployed-engineering/The gap between a working model and a running system, and what a forward-deployed engineer does about it.Wed, 12 Aug 2026 00:00:00 GMThttps://mayurpatil.ai/writing/jev-and-system-one-models/https://mayurpatil.ai/writing/jev-and-system-one-models/What TypeSafe's Jev is, how RLCD calibration works, and where a calibrated decision model fits in enterprise AI.Fri, 09 Oct 2026 00:00:00 GMThttps://mayurpatil.ai/writing/harness-engineering/https://mayurpatil.ai/writing/harness-engineering/What harness engineering means, what the evidence says works, and how it applies to enterprise agents.Sun, 27 Sep 2026 00:00:00 GMThttps://mayurpatil.ai/writing/the-ontology-is-the-product/https://mayurpatil.ai/writing/the-ontology-is-the-product/Why enterprise AI failures are definitional, what an ontology is, and how to build one small enough for a team to own.Thu, 03 Sep 2026 00:00:00 GMThttps://mayurpatil.ai/writing/specification-driven-development/https://mayurpatil.ai/writing/specification-driven-development/How spec-driven development works, what Spec Kit and AI-DLC each add, and the practice that makes either deliver.Fri, 21 Aug 2026 00:00:00 GMThttps://mayurpatil.ai/writing/reinforcement-learning-building-intelligence/https://mayurpatil.ai/writing/reinforcement-learning-building-intelligence/A research-led account of how reinforcement learning changed language models, reasoning and enterprise AI design after 2020.Thu, 18 Jun 2026 00:00:00 GMT