field notes

build logs, tighter lessons

short notes from work on agentic analytics, compounding datasets, and signal extraction across live systems.

project

topic

April 09, 2026

Current Workflow for Data Analytics

Supercharging ML analytics with agentic AI for the 10x data analyst

agentic ai data analytics machine learning codex

menacon regulated-plants

April 08, 2026

We have reached the beginning of the agent's final form, with three core components

agentic ai coding OpenClaw

March 25, 2026

Agents can compress a lot of exploratory work, but they still fail when the problem frame is vague, unstable, or loaded with hidden assumptions.

agentic ai eda workflows

analyse.cefr regulatedplants

March 23, 2026

Public data is useful for prototypes, but repeated product use usually demands a private collection loop that compounds over time.

datasets data moat product

regulatedplants procureplus (menacon)

March 20, 2026

Open agent stacks are improving fast, but most failures still come from weak evaluation and orchestration rather than raw model quality.

agents evaluation orchestration

analyse.cefr

March 17, 2026

For technical work, concrete build notes and operating lessons create more signal than broad product positioning.

writing distribution positioning

procureplus (menacon) regulatedplants analyse.cefr