Tutorials, insights, and lessons from building with AI.
An agent that works once isn't the same as an agent you can trust. This piece breaks down how to build a real evaluation harness — designing a task set that covers more than the happy path, running tasks repeatedly in sandboxed environments, grading with a mix of rules, LLM-as-judge, and human review, and closing the loop by turning every production failure into a permanent regression test.
Read article →
An LLM alone can't act — it's the surrounding harness (control loop, tools, memory management, sandboxing) that turns a model into a working agent. This piece breaks down the core components of agent harnesses, why harness quality often matters as much as the model itself, and the key challenges — from context rot to tool proliferation to long-horizon reliability.
Read article →
AI is making software development dramatically faster by generating code in seconds. But after months of working with AI-assisted development, I've realized that coding was never the biggest bottleneck. The real challenge is understanding the business problem, defining clear requirements, and aligning stakeholders. In the AI era, the engineers who thrive won't be those who write code the fastest—they'll be the ones who ask the best questions.
Read article →
AI can generate algorithms in seconds, but it can't reliably choose the best one for your scale, constraints, and trade-offs. As coding becomes cheaper, DSA shifts from being an implementation skill to the foundation of engineering judgment.
Read article →
AI is changing how marketing gets done. The future belongs to marketers who use AI not just to create content, but to think strategically, understand customers better, and make smarter decisions.
Read article →
AI is changing how we build software. Here is how software engineers should approach it.
Read article →
The future won't belong to the engineers who write code the fastest. It will belong to those who build the most accurate mental model of the problem—and help everyone else converge on the same understanding.
Read article →
A simple conversation during my drive to work made me realize how differently non-technical people perceive AI. When I said AI writes most of my code, the response was simply, "Uchhhh..." 😔 That moment inspired me to explain why AI isn't replacing engineers—it's changing what engineers do. Just like every major technological shift before it, AI is automating the manual work so humans can focus on higher-value thinking and problem-solving.
Read article →Professional help
When your career feels heavy, eight minutes of focused professional support can help you feel heard and less alone. I give 8 minutes every day to anyone struggling in their career.
