Skip to main content

Blog

Thoughts on software engineering, cloud, AI, and building things.

The Missing Instinct: What AI Doesn't Understand About the Real World Featured
technical · 19 min read

The Missing Instinct: What AI Doesn't Understand About the Real World

Artificial Intelligence has reached a point where it can solve complex mathematical problems, generate production-ready code, explain scientific theories, and converse with remarkable fluency. To many people, this feels like intelligence. But beneath the impressive responses lies a fundamental difference between humans and AI—one that rarely receives the attention it deserves. Humans don't make decisions using knowledge alone. We rely on instinct. Fear. Experience. Uncertainty. And millions of years of evolution that taught us one simple objective: Stay alive. Large Language Models have none of these. This essay explores why intelligence without instinct can become surprisingly dangerous, why confidence should never be mistaken for understanding, and why the future of AI depends as much on engineering responsible systems as it does on building better models.

AIArtificial IntelligenceAI SafetyAI Ethics
The Missing Instinct: What AI Doesn't Understand About the Real World
technical · 19 min read · ★ Featured

The Missing Instinct: What AI Doesn't Understand About the Real World

Artificial Intelligence has reached a point where it can solve complex mathematical problems, generate production-ready code, explain scientific theories, and converse with remarkable fluency. To many people, this feels like intelligence. But beneath the impressive responses lies a fundamental difference between humans and AI—one that rarely receives the attention it deserves. Humans don't make decisions using knowledge alone. We rely on instinct. Fear. Experience. Uncertainty. And millions of years of evolution that taught us one simple objective: Stay alive. Large Language Models have none of these. This essay explores why intelligence without instinct can become surprisingly dangerous, why confidence should never be mistaken for understanding, and why the future of AI depends as much on engineering responsible systems as it does on building better models.

AIArtificial IntelligenceAI Safety
The Myth of Token Maxxing
technical · 5 min read

The Myth of Token Maxxing

"Just give the model more context." It's one of the most common pieces of advice you'll hear when building AI applications. Upload the entire codebase. Paste the whole documentation. Use the largest context window available. More tokens must mean better answers... right? Not quite. As context windows continue to grow, many engineers have started treating them like unlimited storage rather than a finite engineering resource. The result is bloated prompts, higher costs, increased latency, and sometimes even worse responses. The real goal isn't to maximize token count. It's to maximize information density. Let's bust the myth of token maxxing.

AILLMPrompt Engineering
Most Organizations Need RAG Before They Need Fine-Tuning
technical · 7 min read

Most Organizations Need RAG Before They Need Fine-Tuning

One of the most common questions I hear from teams starting their AI journey is: "Should we fine-tune our own model?" At first glance, it seems like the obvious solution. If a model doesn't know your business, train it on your company's data. Problem solved. Except that most organizations don't actually have a model problem. They have a knowledge problem. In many enterprise environments, Retrieval-Augmented Generation (RAG) delivers significantly more value than fine-tuning while requiring less complexity, lower cost and faster implementation. In this article, I'll explain why most organizations should start with RAG before considering fine-tuning and how understanding the difference can save months of engineering effort.

AIAgentic AIPlatform Engineering
Why I Wrote Practical AI for Platform and Cloud Engineers
technical · 6 min read

Why I Wrote Practical AI for Platform and Cloud Engineers

Over the past two years, I've watched AI become part of almost every engineering conversation. Architecture reviews suddenly included AI assistants. Platform roadmaps started mentioning LLMs. Developer productivity tools became AI-powered. Organizations began asking how AI could improve operations, security, support and automation. Yet whenever I tried to learn more, I kept encountering the same problem. Most AI content was written for data scientists, machine learning researchers or business leaders. Very little was written for engineers like me. Engineers who build platforms. Engineers who operate Kubernetes. Engineers who manage cloud environments. Engineers who care about reliability, security, observability, and architecture. That realization eventually became the foundation for a book. This is the story behind why I wrote Practical AI for Platform and Cloud Engineers.
Building AgentX: A Multi-Agent AI Platform for Predictive Maintenance and Autonomous Industrial Operations
technical · 7 min read

Building AgentX: A Multi-Agent AI Platform for Predictive Maintenance and Autonomous Industrial Operations

Industrial systems generate massive volumes of telemetry every second, yet many maintenance workflows remain reactive. Engineers often discover problems only after equipment fails, leading to production downtime, increased operational costs, and delayed maintenance cycles. To address this challenge, I designed AgentX, a multi-agent AI platform that combines industrial telemetry, enterprise workflow automation, and autonomous decision-making to enable predictive maintenance at scale. By integrating AI agents, real-time telemetry streams, ServiceNow workflows, and human-in-the-loop governance, AgentX transforms raw machine data into actionable maintenance insights while maintaining enterprise-grade safety and auditability. In this article, I'll walk through the architecture, design decisions, and lessons learned while building AgentX.

Predictive MaintenanceMulti-Agent SystemsServiceNow
Scaling Dentist Channel Online: Lessons from Building a Global Dental Education Platform
technical · 8 min read

Scaling Dentist Channel Online: Lessons from Building a Global Dental Education Platform

Most scaling stories focus on technology. This one starts with a community. Dentist Channel Online (DCO) is a global platform connecting dental professionals through education, webinars, conferences, news and digital learning experiences. Over the years, the platform evolved from a traditional web application into a large-scale ecosystem serving dental professionals across multiple countries and generating millions of requests every month. As traffic grew, so did the engineering challenges. Database bottlenecks, traffic spikes during webinars, asynchronous processing requirements, media delivery and operational reliability all became critical concerns. This article shares the lessons I learned while helping scale Dentist Channel Online, including architectural decisions, performance optimizations and the realities of operating a Laravel application under production workloads.

LaravelPHPBackend Engineering
Self-Regulating Ecosystem in Water Lily Tubs
gardening · 4 min read

Self-Regulating Ecosystem in Water Lily Tubs

One of the biggest concerns when growing water lilies in containers is mosquito breeding. Any standing water can quickly become a breeding ground for mosquito larvae if left unmanaged. Instead of relying on chemical treatments, I decided to establish a natural ecosystem inside my water lily tubs. Over time, this small ecosystem became self-regulating and now requires very little intervention.

Water LilyMosquito ControlGuppies
How I Built Self-Sustaining Water Lily Tubs Using Kitchen Waste and Garden Debris
gardening · 4 min read · ★ Featured

How I Built Self-Sustaining Water Lily Tubs Using Kitchen Waste and Garden Debris

Water lilies are often considered difficult plants that require expensive fertilizers and specialized pond soil. My experience was quite different. I wanted to create a low-cost, nutrient-rich growing medium using materials already available at home. After several weeks of preparation and patience, the result has been rewarding healthy water lilies producing continuous blooms almost every day.

Water LilyTropical Water LilyOrganic Gardening
How We Designed a Self-Healing Cloud System That Can Recover from VM Failures Automatically
technical · 6 min read · ★ Featured

How We Designed a Self-Healing Cloud System That Can Recover from VM Failures Automatically

Traditional monitoring systems are excellent at detecting infrastructure issues, but they still rely heavily on engineers to investigate and resolve incidents. As cloud environments continue to scale, this reactive approach leads to increased downtime, operational overhead, and slower recovery times. In this article, I explore the architecture behind a self-healing cloud platform designed to automatically detect, diagnose, and remediate virtual machine failures. By combining Azure Monitor, Log Analytics, automation workflows and AI-assisted incident analysis, the system aims to significantly reduce Mean Time To Recovery (MTTR) while improving infrastructure reliability.