Skip to content
View angshur's full-sized avatar

Block or report angshur

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
angshur/README.md

Angshuman Rudra

Product leader for data and AI platforms. Former backend and data engineer (Adobe, Yahoo). MBA, Cornell.

What I work on: making AI agents trustworthy on real business data. That comes down to three questions: is the number right, is the agent allowed to do this, and does it still hold when many agents run at once. Each repo below takes one of them, ships with an eval, and lists what it does not prove.

Start here: one definition, every surface

measurement-md defines each business metric once, in a plain file, and serves that definition to every place a number gets produced:

Surface What it does
MCP server Claude, ChatGPT and other agents look up the governed definition before they write SQL
Query gate Refuses an agent's query that breaks a definition and names the clause (62 of 62 planted mistakes caught)
CLI lint, explain, check, export from a terminal or CI
Snowflake Compiles to a semantic view YAML for Cortex Analyst: aliases become synonyms, anti-patterns become SQL-generation rules

The three questions

Question Repo Evidence
Is the number right? measurement-md Linter: 360 of 360 planted defects. Gate: 117 of 117 correct requests allowed
data-layer-interop-checker Finds broken joins in a marketing data stack: 100% planted-defect detection, 0 false alarms
Is the agent allowed to do this? warrant Scoped, revocable authority for agents: 16 of 16 attacks denied for the expected reason
agent-ledger An audit trail of what an agent did and on whose authority
Does it hold at scale? quorum Four agents, no faults, 384 lost updates. With the runtime fixes: 0
Is the platform ready for agents? agent-readiness-scanner Scores a platform on APIs, permissions and docs, with a calibration eval
Did the change help? prompt-bench-cli Golden-set harness with paired statistical comparison and a regression gate

How I build

  • Eval first. Every repo has planted defects and clean controls. A clean control that fails means the harness is wrong, not the code.
  • Limits in the README. Synthetic data says so. Untested assumptions say so.
  • Standard library only where possible, so anyone can run it in a minute.

These are independent projects on synthetic data. · LinkedIn

Popular repositories Loading

  1. web-development web-development Public

    Web Development Projects

  2. angshucornell.github.io angshucornell.github.io Public

    Python

  3. pm-interview-questions pm-interview-questions Public

    Interview Questions for Product Managers

    HTML

  4. backlog-grooming-checklist backlog-grooming-checklist Public

    HTML

  5. competitive-analysis-template competitive-analysis-template Public

    HTML

  6. discovery-questions-product discovery-questions-product Public

    HTML