Essays, papers and the book.
- The Review Room: how one model reads 100 languages
8 Oct 2026
An interactive walkthrough of XLM-RoBERTa, the model that learned to read customer reviews in 100 languages, told through a team that needed it. - Attention Is All You Need… Then Everything Else Too
27 Sept 2026
Prompts, context, tools, agents, workflows, orchestration, harnesses. They're really body parts grown around one brain: the transformer. Scroll, and watch it become a worker. - Attention Is All You Need, explained without the jargon
25 Sept 2026
The 2017 Google paper behind ChatGPT, Gemini and Claude, explained with animated diagrams and everyday examples for readers with no technical background. - From "Domain Agents" to Agent Skills: Why Modular AI is Winning the Enterprise in 2026
20 Mar 2026
In 2025 I argued for modular, plug-and-play expert agents instead of one monolithic model. Agent skills show the industry has caught up, and why it matters for industrial maintenance. - POV #2 — AI Is Already Saving Time: What 40–60 Minutes a Day Means on the Plant Floor
3 Jan 2026
When people hear that AI can save 40–60 minutes per employee per day, the reaction is often mixed. Some think, “That’s exaggerated.” Others worry, “Is this about cutting jobs?”… - POV #1 — Why the State of Enterprise AI in 2025 Matters for Oil & Gas Maintenance
3 Jan 2026
I’ve spent over two decades in oil and gas maintenance, turnarounds and Project management. And if there’s one thing I’ve learned, it’s this: maintenance isn’t just about… - The Cost of Unwanted Intelligence
Book, Out now · 2026
Why enterprises pay for AI they don't use, told through the Intelligence Ledger and the six hidden costs that eat AI returns. - Enterprise Agentic AI: A Harness-First, Role-Responsibility-Driven Architecture
Drishti.ai whitepaper, v1.1 (41 pages) · 2026
A reference architecture for running AI agents in the enterprise: deterministic scripts do the heavy lifting, models do the judgment, and every agent has a defined role, responsibility and supervisor. - Why AI Adoption Is Still Low in enterprises
4 Dec 2025
I still remember those long bus rides through northern Uttar Pradesh — the Babrala ( a village in northern India) to Agra route on UP Roadways. Dusty roads, blaring horns, chai… - Why “Replace the Professional” theory fails for AI adoption in industries
21 Nov 2025
Introduction: The Paradox of AI and Human Expertise A familiar story in every industrial plant: a massive machine suddenly stops. Operators check gauges, valves, and control… - Is Your RAG System Built on a House of Cards? A First-Principles Guide to Enterprise AI
15 Sept 2025
Many enterprise RAG projects are brittle and untrusted. Three first-principles pillars, Relevance, Reliability and Responsibility, for building one that lasts. - Enhancing Enterprise Data Collection through Interactive Template-Filling Agents with Localized Language Models
June 2025International Journal of Scientific Development and Research 10(2):b45-b48, Published · Jun 2025
Structured user data gathering in enterprise environments is critical but sometimes error-prone and boring when using traditional static forms. We propose an innovative interactive template-filling agent solution based on localized compact language models (LLMs). These agents guide users through dynamic prompts, context recall, and instant feedback to complete filled-in data form templates. We show a proof-of-concept deployment of this system with a Django-based front end and a locally run LLM (e.g., Flan-T5 or LLaMA2). Comparative evaluation with static ones on an HR onboarding task revealed statistically significant gains in data completeness (95% vs. 75%), accuracy (6% vs. 18% error rate), and user satisfaction (4.4/5 vs. 3.2/5). Our solution facilitates effective, privacy-aware data gathering, applicable in HR, legal, and customer onboarding scenarios. - Email Intelligence: An AI-Driven Framework for Context-Aware office Communication Management
February 2025International Journal of Scientific Development and Research 10(2):b45-b48, Published · Feb 2025
This paper presents an innovative, AI-based intervention that revolutionizes email communication. We propose an intelligent system for the processing of email chains among multiple stakeholders by using deep natural language processing and machine learning techniques to present a more intuitive and efficient interface of email understanding and ingesting. The AI Tool extracts vital information, finds critical topics in communication threads, and generates substantial summaries while the stakeholder expectation and intentions are held intact. Solutions offered by our system specifically address the expanding challenges of information overload in professional communications by creating an abstract representation of email chains that distill essential points without compromising their clarity or context-a method that keeps knowledge transfer with inclusive communication among all participants. The research contributes to the field by developing a new framework that analyzes communication patterns and at the same - Domain Subject Matter Expert-Driven AI/ML Implementation Framework
International Journal of Scientific Development and Research (IJSDR), Vol. 10, Issue 2, Published · Feb 2025
Argues that the people who know the work best should lead its automation, and sets out a framework for experts to design, build and supervise AI/ML solutions with IT as enabler rather than gatekeeper.
Nothing here yet.
Say hello
Always happy to swap notes on AI in industry, maintenance and reliability.
The Plant Floor
AI news, tested against reality. One short read every Sunday.