Writing

Case studies and technical essays on AI engineering, agents, and production data

Technical writing on AI-agent systems, Text-to-SQL, coding agents, evaluation harnesses, data platform work, and time series ML.

Longer-form case studies and technical essays on AI-agent systems, evaluation, and production data platform work live here.

Shorter notes and experiments live in TIL / Random Notes.

A Clean Zero Is the Most Dangerous Answer

Text-to-SQL

LLM Evaluation

Coding Agents

Full-Stack

On a real ERP, the hard part of text-to-SQL isn’t writing SQL. It’s retrieval, grounding, and knowing when a query that ran cleanly is silently wrong. Field notes, plus the open-source rebuild.

Jun 28, 2026

13 min

What Counts as Evidence.

Software Engineering

Coding Agents

Testing

A practitioner’s reference for migration testing. Five layers of tests, why every artifact carries a cryptographic hash, and how to sequence the suite so the oracle leads the implementation.

May 2, 2026

18 min

Fast Wrong Is Worse Than Slow Right

Data Engineering

Performance Optimization

Snowflake

Coding Agents

The fastest optimization I tested was wrong. The working redesign used multi-scale benchmarks and row-level validation to deliver 50%+ compute savings and 5-10x faster daily runs.

Apr 7, 2026

6 min

How I Validated a Large Data Platform Migration in a Week with an AI Coding Agent

Artificial Intelligence

Data Engineering

Coding Agents

How I used an AI coding agent to build a cross-platform validation workflow and dashboard for a large Databricks-to-Snowflake migration.

Apr 1, 2026

11 min

Don’t Port the Syntax. Port the Evidence.

Artificial Intelligence

Software Engineering

Coding Agents

A practitioner’s guide to large code migrations and clean-room-style rewrites with coding agents, focused on evidence, behavioral oracles, and verification discipline.

Mar 9, 2026

10 min