It Reads Better Than It Decides

A chatbot that invents a fact says it to one person, once. A connector that invents a fact writes it into the graph, where it looks exactly like the true ones. So I let the model read, and took away its ability to decide.

September 30, 2026 · 15 min

We've Seen This Movie Before

Prompt attacks are a new kind of threat. The tooling growing up around them is a ladder security has already climbed once.

August 24, 2026 · 14 min

Cozy Bear, The Dukes, and APT29 Walk Into a Database

One hacking group can carry half a dozen names. Merging them is easy until you merge two different groups by accident — which I did, ten times, including two pairs of state bodies from different countries.

May 7, 2026 · 14 min

Teaching a Chatbot to Read a Spider Web

A knowledge graph holds its best facts in the connections between records. Vector search can only read what is written down. So I wrote the connections down.

April 9, 2026 · 14 min

From Half a Day to a Coffee Break

Keeping a knowledge base current in minutes instead of half a day, by never rebuilding what has not changed.

March 12, 2026 · 14 min

I'm Ashwin. I build the engineering that lets organizations get ahead of attacks, and write about the decisions behind it.

Email · Resume · GitHub · LinkedIn