Hi there 👋

Thanks for stopping by—I’m Devendra, an ML engineer who ships systems that have to run reliably in production, not just score well in a notebook. I write here about edge perception, MLOps, and private Gen AI. For work, I take on freelance engagements—from AI strategy and “do we need this?” assessments to MLOps audits and a practical path to integrate ML or Gen AI when it actually fits. See consulting for offerings, CV for background, or reach out via the links below.

AI Governance, From First Principles

Table of Contents 1. Where I started 2. Why, what, and how 3. From two questions to a map 4. Reading the frameworks 4a. The EU AI Act 4b. NIST’s AI Risk Management Framework 4c. The OECD AI Principles 4d. The UK’s pro-innovation approach 4e. Frontier (scaling-law) governance 5. What the map shows 6. Where it’s going References, by framework 1. Where I started When I first tried to read about AI governance, I found myself in an alphabet soup. There’s a pile of names to learn: the EU AI Act, NIST, Bletchley, the OECD Principles, Korea’s Basic Act, “the omnibus,” Singapore’s framework, the UK’s different one, a treaty and the abbreviations stack up faster than the understanding does. If you’ve tried, you’ve probably hit the same wall. ...

The Last Mile Is the Whole Job

Every parcel you’ve ever received traveled most of its distance cheaply. A container ship crosses an ocean for cents per kilogram. A truck runs a highway lane at near-perfect utilization. Then the parcel reaches a depot twenty minutes from your door, and the economics fall off a cliff. Figure 0: The Last Mile Delivery Problem The numbers say it plainly. Last-mile delivery now absorbs roughly 53% of total shipping cost — up from 41% in 2018, and now the single largest line item in the fulfillment chain, having quietly overtaken warehousing and middle-mile freight combined [1][2]. ...