DEV Community

Ricardo Feldhaus Krisanoski profile picture

Ricardo Feldhaus Krisanoski

I build AI-assisted operating systems for CRM, WhatsApp sales, revenue operations and internal workflows.

Joined Joined on  github website
A Public Case Study Should Be an Evidence Map

A Public Case Study Should Be an Evidence Map

Comments
3 min read
What a Small RAG Fine-Tuning Experiment Can Actually Prove

What a Small RAG Fine-Tuning Experiment Can Actually Prove

Comments 1
6 min read
Your AI Workflow Needs an Exception Contract

Your AI Workflow Needs an Exception Contract

Comments 1
3 min read
The Cost of Changing Your Automation Is Part of Its Design

The Cost of Changing Your Automation Is Part of Its Design

Comments 1
3 min read
A Retrieval Result Needs an Allowed-Use Field

A Retrieval Result Needs an Allowed-Use Field

Comments 1
3 min read
A Read-Only Integration Is a Deliberate Product Boundary

A Read-Only Integration Is a Deliberate Product Boundary

Comments
3 min read
Make Handoffs Explainable Before You Make Them Automatic

Make Handoffs Explainable Before You Make Them Automatic

Comments
3 min read
Freeze the Comparison, Not the Working Document

Freeze the Comparison, Not the Working Document

Comments
3 min read
What a Production Brief Should Say When the Data Is Incomplete

What a Production Brief Should Say When the Data Is Incomplete

Comments
3 min read
Read Access and Action Authority Belong to Different Contracts

Read Access and Action Authority Belong to Different Contracts

Comments
3 min read
Do Not Turn an Observability Failure into a Business Outage

Do Not Turn an Observability Failure into a Business Outage

Comments
3 min read
A Retry Needs an Identity, Not Just Another Attempt

A Retry Needs an Identity, Not Just Another Attempt

Comments
3 min read
An Approval Queue Is a Product Interface

An Approval Queue Is a Product Interface

Comments
3 min read
A Quote Comparison Needs an Equivalence Policy

A Quote Comparison Needs an Equivalence Policy

Comments
3 min read
A Useful Context Packet Should Explain What It Left Out

A Useful Context Packet Should Explain What It Left Out

Comments
3 min read
An Evaluation Score Is Only as Useful as Its Target

An Evaluation Score Is Only as Useful as Its Target

Comments
5 min read
A Daily Production Brief Is a Coordination Interface, Not a Dashboard

A Daily Production Brief Is a Coordination Interface, Not a Dashboard

Comments
4 min read
Operational Knowledge Is a Product Surface, Not a Folder

Operational Knowledge Is a Product Surface, Not a Folder

Comments
4 min read
What I log from AI workflows before I trust the automation

What I log from AI workflows before I trust the automation

Comments 1
3 min read
AI Coding Agents Need a Decision Gate Before the Implementation Loop

AI Coding Agents Need a Decision Gate Before the Implementation Loop

Comments
5 min read
Why my first RAG layer starts in Postgres, not in a standalone vector database

Why my first RAG layer starts in Postgres, not in a standalone vector database

Comments
3 min read
I added a context resolver before an AI sales agent replies

I added a context resolver before an AI sales agent replies

2
Comments
3 min read
loading...