Welcome to the notes
A résumé tells you what I did. These notes are for the part that doesn't fit on a résumé - the why, the wrong turns, and the long stretches of the trail.
This site has two pages doing very different jobs. The homepage is a pitch, in the best sense: here is the research, here is the work, here is the trajectory through the latent space. The architecture page is a drawing sheet - fifteen years of estates nobody wanted to touch, turned into things that ship.
Both are summaries. The notes are the rest. A couple of reasons I wanted them.
Science is slow, and papers hide that
Take the DOA framework - Dynamic Oracle Adaptation. The published line is one sentence: a quantum oracle that adapts instead of staying fixed, trading a little coherence for a lot of flexibility. The IEEE paper is the formal version. But between "what a fixed oracle can't do" and "here is the adaptive construction" there was a long, mostly unglamorous middle: simulations that contradicted the intuition, notation I had to rebuild twice, and the realization that the hardest problem wasn't the math - it was the vocabulary. Nobody had named this thing, because nobody had built it. We had to invent the language as we went.
I don't think the papers should carry that. Papers should be tight. The notes can be honest about the path.
Two audiences, one voice
I work a strange seam. By weekday I'm a Solution Architect - COBOL estates, mainframe migrations, GenAI rewriting code it was never trained to understand. By weekend I'm a PhD researcher at IIIT Nagpur, drawing quantum circuits. The questions that interest me sit exactly on that seam:
- What does a mainframe migration and a quantum oracle have in common? (More than you'd think - both are state that has to keep working while you change the system around it.)
- Where does an LLM, which predicts the next token, meet a quantum state, which doesn't decide until measured? Both are machines for managing uncertainty under pressure.
- How do you explain any of this to someone who has to sign a budget for it?
What this space is for
Three kinds of posts, roughly:
- Explainers - a fixed oracle vs. an adaptive one; the coherence budget; what "measure and let the oracle learn" actually means on a NISQ device.
- Build logs - MindSpark and mindspark-mcp: the decisions that a repo's README flattens away.
- Field notes - migration tradecraft, the DOA vocabulary problem, the occasional opinion.
"Depth gives you power, breadth gives you reach. Wisdom is the knowing when to use which."
- the working motto of this whole site
If you're reading via RSS, welcome. If you arrived from the homepage, the two posts that follow unpack the core of the research - what an adaptive oracle actually changes, and the coherence budget between HQ-DOA and FQ-DOA. Comments are best sent to email; the notes are static, and so is my inbox. That's the point.