🔍 Deep Dive: Pragmatic AI Tools Series' Articles

Cover image for Your Flaky Tests Aren't Flaky

Debashish Ghosal

Jul 18

Your Flaky Tests Aren't Flaky

#discuss #ai #python #testing

6 min read

Cover image for I Planned 10 LLM Evaluation Experiments And Only Ran 1. It Was Enough.

Debashish Ghosal

Jul 26

I Planned 10 LLM Evaluation Experiments And Only Ran 1. It Was Enough.

#discuss #ai #webdev #llm

5

2

13 min read

Cover image for I Thought Building Agent Observability Was a Detector Problem. I Was Wrong.

Real production data broke 35 detectors instantly

Debashish Ghosal

Aug 7

I Thought Building Agent Observability Was a Detector Problem. I Was Wrong.

#ai #opentelemetry #llmops #opensource

27

19

8 min read

Cover image for I Built Scenario Packs for Agent Regression Testing. The Integration, Not the Judge, Broke Me.

Adapters need conformance suites too

Debashish Ghosal

Aug 8

I Built Scenario Packs for Agent Regression Testing. The Integration, Not the Judge, Broke Me.

#discuss #ai #agentskills

18

21

15 min read

Cover image for I Stopped Trusting AI Agents With Tools. So I Built a Gatekeeper.

83 real agents tested zero mocks

Debashish Ghosal

Aug 13

I Stopped Trusting AI Agents With Tools. So I Built a Gatekeeper.

#ai #agents #security #gatekeeper

55

64

14 min read

Cover image for I Shipped an Agent Gatekeeper (v0.1). 14 Developers Showed Me What I Missed. Here's v0.2 — a Control Plane.

Debashish Ghosal

Aug 15

I Shipped an Agent Gatekeeper (v0.1). 14 Developers Showed Me What I Missed. Here's v0.2 — a Control Plane.

#ai #agents #gatekeeper #llm

8

2

13 min read

Cover image for 9 Bugs That All Looked Like a Working System

Silent statistical traps in self-improving loops

Debashish Ghosal

Sep 1

9 Bugs That All Looked Like a Working System

#ai #agents #llm #correcting

20

14

10 min read

Cover image for I Built an AI That Rewrites Its Own Prompts — Its Safety Gate Rejected Every Single Edit

Debashish Ghosal

Sep 1

I Built an AI That Rewrites Its Own Prompts — Its Safety Gate Rejected Every Single Edit

#ai #promptengineering #llm #agents

24

7

8 min read

Cover image for I Let an LLM Rewrite Its Own Prompt. The Real Win Was the Gate That Rejected It.

Debashish Ghosal

Sep 2

I Let an LLM Rewrite Its Own Prompt. The Real Win Was the Gate That Rejected It.

#healthydebate #ai #llm #agents

11

1

6 min read

Cover image for My Agent Found Real Improvements. The Statistics Still Killed the Promotion.

Debashish Ghosal

Sep 2

My Agent Found Real Improvements. The Statistics Still Killed the Promotion.

#healthydebate #ai #agents #llm

11

3

6 min read

Cover image for I Tried 4 Models to Save My Self-Improving Agent. All 4 Failed.

Debashish Ghosal

Sep 3

I Tried 4 Models to Save My Self-Improving Agent. All 4 Failed.

#healthydebate #ai #llm #agents

24

2

7 min read

Cover image for When Your Benchmark Finally Tells the Truth

Debashish Ghosal

Sep 6

When Your Benchmark Finally Tells the Truth

#healthydebate #ai #llm #agents

13

6

8 min read

Cover image for We Could Have Shipped on Local Models Alone

Debashish Ghosal

Sep 6

We Could Have Shipped on Local Models Alone

#healthydebate #ai #llm #testing

6

2

12 min read

Cover image for A Better Model Improved the Numbers. It Didn't Fix the Product.

Debashish Ghosal

Sep 7

A Better Model Improved the Numbers. It Didn't Fix the Product.

#ai #llm #testing #programming

15

4

10 min read

Cover image for The 6-Line Fix That Outperformed My Entire Matcher Week

Debashish Ghosal

Sep 8

The 6-Line Fix That Outperformed My Entire Matcher Week

#ai #llm #agents #testing

18

5

11 min read

Cover image for My 3B Model Found a Shortcut. It Took Me Three Fixes to Close It.

Debashish Ghosal

Sep 8

My 3B Model Found a Shortcut. It Took Me Three Fixes to Close It.

#ai #llm #programming #testing

14

2

9 min read

Cover image for I Tried to Poison My Agent's Rule Store. It Produced 20 Triggers. Zero Got In.

Debashish Ghosal

Sep 9

I Tried to Poison My Agent's Rule Store. It Produced 20 Triggers. Zero Got In.

#ai #security #llm #testing

13

1

9 min read

Cover image for A Rule Can Be Specific and Still Be Too Broad

Debashish Ghosal

Sep 9

A Rule Can Be Specific and Still Be Too Broad

#ai #llm #broad #testing

12

3

10 min read

Cover image for I Shipped a Fix That Fixed Nothing. Here's Why I Kept It.

Debashish Ghosal

Sep 10

I Shipped a Fix That Fixed Nothing. Here's Why I Kept It.

#ai #llm #testing #programming

20

3

10 min read

Cover image for The Model Wrote the Right Rule and My Replay Rejected It: The Extraction-vs-Replay Split

Debashish Ghosal

Sep 12

The Model Wrote the Right Rule and My Replay Rejected It: The Extraction-vs-Replay Split

#ai #replay #agents #testing

17

7

6 min read

Cover image for My Extraction Score Was 0.08 and the Model Was Innocent: Rebuilding the Ruler

Debashish Ghosal

Sep 13

My Extraction Score Was 0.08 and the Model Was Innocent: Rebuilding the Ruler

#ai #llm #measurement #agents

15

3

8 min read

Cover image for 0/60 Wasn't the Model: The Empty Haystack Behind My Two Worst Corpora

Debashish Ghosal

Sep 14

0/60 Wasn't the Model: The Empty Haystack Behind My Two Worst Corpora

#ai #llm #agents #testing

19

7

7 min read

Cover image for Killed by the Word 'git': One Token of Coincidence, 40 Points of Pass Rate

Debashish Ghosal

Sep 14

Killed by the Word 'git': One Token of Coincidence, 40 Points of Pass Rate

#ai #llm #agents #testing

18

9

8 min read

Cover image for A Floor of 0.80 and a Ceiling of 0.63: The Semantic Channel That Never Fired

Debashish Ghosal

Sep 15

A Floor of 0.80 and a Ceiling of 0.63: The Semantic Channel That Never Fired

#cli #github #opensource #python

17

2

7 min read

Cover image for The 15-Line Test That Catches the #1 Killer of Operator Trust

Debashish Ghosal

Oct 4

The 15-Line Test That Catches the #1 Killer of Operator Trust

#healthydebate #ai #testing #llm

24

8

6 min read