Writing

Start here for the practical arc: AI diligence, architecture judgment, cost discipline, and operating cadence decide whether ambitious systems survive reality.

Flagship diligence essay AI Diligence Is Becoming Technical Diligence A practical framework for evaluating the whole AI system: workflow, evals, security, cost, governance, and defensibility. 9 min Canonical essay The Enterprise AI Bottleneck Is Not the Model Why enterprise AI value often stalls in workflow ownership, trust boundaries, integration cost, and executive operating cadence. 14 min Operator economics The Real Cost Of Just Add AI Inference is only the visible part. Production AI costs include retries, latency, review, support, security, and margin design. 8 min CTO office Architecture Reviews That Do Not Become Theater A practical review loop for decisions, evidence, tradeoffs, owners, and follow-up. 7 min

The clusters map recurring decision surfaces across systems, leadership, diligence, and builder economics.

• AI Diligence Is Becoming Technical Diligence9 min
• What Investors Should Ask Before Funding An AI Company8 min
• The Real Cost Of Just Add AI8 min
• Architecture Reviews That Do Not Become Theater7 min
• What I Would Teach A First-Time CTO In 90 Days8 min
• The Enterprise AI Bottleneck Is Not the Model14 min
• AI Safety Is Not a Checkbox: Building Guardrails That Actually Hold11 min
• Multi-Agent Orchestration Patterns That Survived Production9 min
• From Copilot to Cursor to Claude Code: An Honest Migration Log11 min
• Red-Teaming Our Own AI: An Enterprise Safety Playbook6 min
• Small Teams, Big Models: Why 15 Engineers Now Ship Like 2009 min
• Claude Code Changed How My Team Writes Software10 min
• The LLM Security Audit Nobody Wants to Do9 min
• What Broke When We Let Agents Talk to Each Other7 min
• Twelve Engineers, Three Agents, One Platform12 min
• Beyond the LLM Monolith: How Enterprises Capture AI Value7 min
• My Experience with Frontier LLMs: A Comparative Study6 min
• The Agentic Shift: Building AI That Does, Not Just Suggests10 min
• Lessons from Enterprise AI Projects9 min
• Technical Debt in the Age of AI-Assisted Coding6 min
• Observability for LLM-Powered Systems8 min
• The Director's Playbook: Running Engineering at Scale9 min
• Why I Stopped Chasing the Latest Model5 min
• Prompt Engineering is Dead, Long Live System Design5 min
• The Architecture of Trust: Designing Systems Humans Can Audit7 min
• Building for India: Latency, Cost, and Regional Nuance8 min
• Engineering Leadership: Diagnostics, Ambition, and Mentorship4 min
• How I Actually Evaluate LLMs for Production8 min
• Designing Dependable Distributed Systems11 min
• Repository Hygiene for Distributed Infra Teams4 min
• My Developer Environment Blueprint5 min