1 min read
Fabric Data Factory Notes: designing pipelines for failure, not the happy path
I focused on making delivery decisions auditable and repeatable—documenting intent, success criteria, and rollback paths to reduce tribal knowledge.
Engineering
26 articles
1 min read
Orchestration Lessons in Fabric: debugging pipeline latency before it becomes a fire drill
I worked on smoothing the handoff between data engineering and AI teams—standardizing feature contracts, embedding validation, and adding lightweight…
1 min read
Orchestration Lessons in Fabric: how I keep orchestration readable over time
I turned implicit processes into explicit operating rules—defining owners, acceptance tests, and lightweight runbooks so teams can move confidently and…
1 min read
Orchestration Lessons in Fabric: debugging pipeline latency before it becomes a fire drill
I focused on making delivery decisions auditable and repeatable—documenting intent, success criteria, and rollback paths to reduce tribal knowledge.
1 min read
Orchestration Lessons in Fabric: designing pipelines for failure, not the happy path
I spent the day reducing cognitive overhead for engineers and analysts—introducing clearer table contracts, simpler failure modes, and concise runbooks that…
1 min read
Data Movement Without Drama: designing pipelines for failure, not the happy path
I spent the day reducing cognitive overhead for engineers and analysts—introducing clearer table contracts, simpler failure modes, and concise runbooks that…
1 min read
Pipelines I Trust in Fabric: reducing brittle dependencies in daily loads
I worked on smoothing the handoff between data engineering and AI teams—standardizing feature contracts, embedding validation, and adding lightweight…
3 min read
The Principal Engineer Question
Someone asked me last week if I miss being a principal engineer. I had to think about it longer than I expected.
3 min read
Real-Time Intelligence in Microsoft Fabric: A Working Example
Here's what it is, how it works, and a pattern that's running in production. Traditional analytics is batch-oriented. Data lands in storage, gets processed…
3 min read
The Code Review That Rewired How I Think
Seven years ago, a senior engineer reviewed my code and said nothing about the code. I've thought about that review almost every week since.
3 min read
Azure AI Foundry: What It Is and Why It Matters
A unified platform for building, deploying, and managing AI applications and agents at enterprise scale.
3 min read
Fine-Tuning vs RAG: The 2026 Decision Guide
In 2024, the community consensus was "RAG first, fine-tune never." In 2026, it's more nuanced.
2 min read
What My Son Taught Me About Debugging
My son Andriel is nine. Last weekend he was building a LEGO set and got stuck. I watched him work through it for twenty minutes without knowing he was…
2 min read
Delta Lake in Microsoft Fabric: Why the Format Matters
When I tell clients "everything in Fabric uses Delta Lake format," the room divides. Data engineers nod. Everyone else says "what?"
2 min read
Prompt Caching: The Performance Win Nobody Talks About
When you send the same prompt prefix repeatedly—system instructions, context documents, examples—the model recomputes them every time. Prompt caching stores…
2 min read
AI Agents vs Workflows: Choosing the Right Pattern
Workflow: Predefined steps. AI handles specific tasks within a fixed pipeline. Deterministic flow.
2 min read
Microsoft Fabric: Six Months In Production
It worked, but managing five services with five billing models and five sets of credentials was painful.
2 min read
Evaluating LLM Outputs: Beyond Vibes
Production AI needs real evaluation. Here's how I approach it. You test with 5 prompts. They look good. Ship it.
2 min read
Data Pipelines: What Production Actually Teaches You
I've built data pipelines for 8 years. Every production deployment humbled me. Here's what the tutorials don't teach.
2 min read
The Real Cost of Context Switching
Last Tuesday I worked 10 hours. Got almost nothing done. Checked my calendar: 6 meetings scattered throughout the day. None longer than 30 minutes. Plenty…
1 min read
Building RAG Systems That Actually Work
This works for demos. It fails in production. Chunking matters more than you think. Random 500-token chunks lose context. A sentence about "the system"…
1 min read
Keeping Deployments Simple
For an app with 100 users. Push to main. App deploys. Done. Don't add any of this on day one.
2 min read
Learning AI as a Traditional Engineer
Three years ago, I was a cloud/data engineer. AI was "something other people do." Now it's a core part of my work.
1 min read
Building Software I Want to Maintain
I'm tired of clever code. I want boring, maintainable code. Early career: "Look at this elegant one-liner!"
1 min read
Testing AI Systems: What Actually Works
Traditional testing assumes deterministic behavior. AI systems are probabilistic. Same input, different output.
1 min read
Complex Problem Solving with o1: Practical Patterns
o1's reasoning capabilities open up new possibilities for tackling complex problems that were previously difficult for AI to handle reliably.