· Ruby Jha · engineering-leadership
When Standups Feel Like Interrogations
How to diagnose whether tight oversight is a trust problem or a legitimate need, and how to hand back autonomy without losing accountability.
The Blog
On building AI systems, leading engineering teams, and the decisions behind both.
- · Ruby Jha · engineering-leadership
Your Team Is Doing the Work. Someone Else Is Taking the Credit.
How to fix invisible attribution in distributed teams, and why credit theft is the most corrosive trust pattern a manager can inherit.
- · Ruby Jha · engineering-leadership
The Reorg Nobody Told Your Team About
How to rebuild trust when your team learned about their own reorg from an org chart update, not a conversation.
- · Ruby Jha · project-deep-dives
LoRA Hit 96% of Full Fine-Tuning. The Default Learning Rate Almost Killed It.
I fine-tuned all-MiniLM-L6-v2 on dating profiles, flipped Spearman from -0.22 to +0.85, and found LoRA hit 96.2% of that with 0.32% of parameters.
- · Ruby Jha · project-deep-dives
I Tested 16 RAG Configs So You Don't Have To: Embedding Choice Matters More Than Chunk Size
Grid search across 16 RAG configurations reveals embedding model selection drives 26% more retrieval quality than chunk tuning.
- · Ruby Jha · engineering-leadership
How a One-Month Fix Saved $250K/Month in Ops Costs
How I scoped a 1-year architecture proposal down to a 1-month solution that cut support calls by 60% and saved $250K in recurring ops costs.
- · Ruby Jha · project-deep-dives
How I Calibrated an LLM Judge That Approved Everything
My first LLM judge had a 0% failure rate. That meant it was useless. This is the story of calibrating it to actually catch failures, and building a correction loop that took synthetic data failures from 36 to zero.
- · Ruby Jha · project-deep-dives
Building 9 AI Projects (While Working Full-Time)
Why I am building 9 AI systems from scratch while working full-time as an Engineering Manager. The portfolio, the progression, and what I have learned so far.