When Prompt Engineering Wasn’t Enough The model was smart.The prompts were detailed.The outputs were… inconsistent.Sometimes it answered perfectly.Sometimes it ignored instructions it followed just one request earlier.We added more examples.We refined prompts.We layered system messages.Eventually, it became clear: this wasn’t…
Continue Reading
Installing VS Code extensions is usually effortless—click Install, wait a second, and move on.Until the day you can’t.Offline machines.Air-gapped servers.Strict enterprise networks.Version pinning requirements.That’s when you realize you don’t just need the Python extension—you need the VSIX file itself.Manual ms-python.python…
The Dashboard Was Green. The Problem Wasn’t. The model passed every metric.Accuracy was high. Precision looked solid. Latency stayed well within limits.The deployment dashboard was glowing green.A week later, a complaint landed in my inbox: “Why was my application rejected…
The Incident That Didn’t Look Like a Bug The dashboard was green.Latency was fine.Accuracy was high.No exceptions. No alerts. Then a manager asked a simple question that froze the room: “Why did the model reject this customer?” We had technical…
The Meeting That Changed How I Looked at Azure AI I used to think “Azure AI” meant one or two smart APIs.Then came a project review where someone casually said: “We’re using Vision for product images, Speech for call transcription,…
The Decision That Looked Correct on Paper Why responsible AI is important didn’t become clear to me after reading a regulation or attending a conference.It became clear after a meeting that felt uncomfortable.A model we deployed had rejected an application.The…
The Experiment That Looked Right—but Wasn’t Best Python package manager for data science wasn’t something I searched for after reading blog posts.I searched for it after a result I trusted turned out to be wrong.The notebook ran.The model trained.The numbers…
The Day My Environment Became the Problem The difference between Poetry and Conda became clear to me the day my Python environment stopped being “just Python.”For a long time, I thought my problems were about code.When something broke, I assumed…
The Decision I Thought Didn’t Matter Poetry vs Pipenv wasn’t supposed to be a big decision.I wasn’t choosing a database.I wasn’t redesigning architecture.I was “just picking a dependency manager.”So I did what most people do.I picked one, told myself I’d…
The Day My CI Failed for No Good Reason How to use uv as a pipx replacement wasn’t a question I asked out of curiosity. It was a question I asked out of frustration. A CI pipeline that had been…
The Notebook That Worked—Until It Didn’t PIP vs CONDA for data science wasn’t a debate I planned to have.It started with a Jupyter notebook that ran perfectly on my laptop.The plots rendered.The model trained.The metrics looked reasonable.Then I sent the…
AWS Cognito to Azure AD B2C migration became inevitable for us when enterprise security requirements started blocking real business deals. I need to have a conversation with myself about the day we lost a major enterprise contract because of a…
When Installs Became the Loudest Part of My Workflow Why uv is faster than pip only became obvious to me when installs stopped being background noise.Not because they failed.Not because they threw errors.But because they interrupted me.I’d write a few…