> Blog
Thoughts on AI, engineering, and building products.
My Problem-Solving Playbook for Post-Silicon Failures
When a chip fails and millions of dollars are on the line, panic is not a strategy. Here is the structured approach I use to go from "something is wrong" to root cause.
Building Real-Time Production Dashboards That People Actually Use
Most engineering dashboards get ignored. Here is what I learned building PowerBI dashboards for semiconductor mass production that stakeholders check daily.
Debugging Under Pressure: Five Strategies That Keep Me Effective When Everything Is on Fire
High-stakes troubleshooting is a mental game as much as a technical one. Here are the strategies I rely on when the clock is ticking and the answer is not obvious.
From Reactive to Predictive: Applying ML to Silicon Yield Data
How we moved from waiting for batches to fail to predicting yield excursions before they happen, using gradient-boosted models on historical parametric test data.
Data Visualization as Storytelling: How I Turn Yield Data Into Decisions
A chart is not a visualization. A visualization tells a story that drives action. Here is how I design dashboards and reports that change what people do, not just what they see.
Using LLMs to Query a Decade of Silicon Failure Reports
We pointed a retrieval-augmented LLM at ten years of root-cause analysis reports and turned tribal knowledge into a searchable, queryable system.
The Wafer Map Does Not Lie: Visual Pattern Recognition in Silicon Debug
Before you run a single statistical test, look at the wafer map. Spatial patterns in silicon data reveal root causes that numbers alone cannot.
The Post-Silicon Mindset: Why Hardware Debugging Is a Software Problem Now
Modern post-silicon validation lives at the intersection of hardware hands-on work and software-driven analysis. Here is how the role has evolved.