Skip to main content

9 AI Projects

4 posts published

  1. 1
    Building 9 AI Projects (While Working Full-Time)
    3 min read

    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.

  2. 2
    How I Calibrated an LLM Judge That Approved Everything
    10 min read

    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.

  3. 3
    I Tested 16 RAG Configs So You Don't Have To: Embedding Choice Matters More Than Chunk Size
    9 min read

    Grid search across 16 RAG configurations reveals embedding model selection drives 26% more retrieval quality than chunk tuning.

  4. 4
    LoRA Hit 96% of Full Fine-Tuning. The Default Learning Rate Almost Killed It.
    8 min read

    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.