Ednalyn C. De Dios

Category: Notes

  • Who Owns the Decision When Automation Acts?

    Automation moves decisions faster and farther than people ever could. That efficiency creates a new question, one that many organizations don’t fully answer until something goes wrong: who owns the decision when automation acts on the organization’s behalf? Ownership matters not because automation fails often, but because when it does, the impact is immediate, visible,…

  • Human-in-the-Loop Is Not Governance

    “Human-in-the-loop” is often presented as the answer to AI risk. It sounds reassuring. It suggests oversight, judgment, and control. In reality, it is usually a placeholder for decisions the organization has not yet fully defined. Having a human involved does not automatically create governance. Governance exists only when authority, responsibility, and accountability are explicit. 1.…

  • What “Safe to Automate” Actually Means

    A practical way to define automation boundaries Most organizations don’t have a hard time deciding whether they should automate. Rather, they struggle with where automation should stop. “Safe to automate” is often treated as a technical question: model accuracy, confidence scores, test coverage. In practice, it’s an operational one. The real challenge is not deciding…

  • Your AI Didn’t Fail. Your Process Did.

    AI exposes operation gaps faster than any audit could. Intro When an AI system produces unexpected outcomes, the first reaction is often to question the technology. But that reaction misses the point. AI doesn’t work in isolation. It executes the processes, rules, thresholds, and decision paths it’s given. When results disappoint, it’s usually because those…

  • When Things Stop Being Simple –

    And what tends to happen next. Most systems make sense at the beginning. A requirements document here, a flowchart there. A diagram shows how the pieces fit together. Inputs go in, outputs come out, and the logic in between makes sense. You can explain it cleanly and feel confident that the explanation will hold. For…

  • CDC Outbreaks

    I was on the hunt for data in the wild and stumbled upon CDC’s open datasets. I noticed that they used a Power BI dashboard. I got inspired and decided to recreate/reverse-engineer the dashboard. Here’s the result: NORS Dashboard The repo can be found at: CDC Outbreaks

  • Enter LangGraph

    I finally bit the bullet and tried out LangGraph. LangGraph is a library for building stateful, multi-actor applications with LLMs, used to create agent and multi-agent workflows. I created an LLM application that facilitates learning by generating questions and answers related to the user’s topic. Github repo

  • From DataFrame to Knowledge Graph

    Visualize the relationships of biblical figures in a knowledge graph using networkx. Github repo

  • Power BI – Labor Optimization Tool

    A Power BI dashboard comparing actual demand versus labor allocation. Github repo

  • Structured Data to Knowledge Graph for Retrieval-Augemented Generation

    Building a knowledge graph-powered RAG from structured data. Github repo

  • Fallout

    Pardon the look and feel of the website right now if they hurt your eyes. I recently started playing Fallout again, and I wanted to create something similar to the Pipboy. Lollllz.

  • Retrieval-augmented generation powered by a knowledge graph

    Automated knowledge graph construction and RAG implementation using OpenAI’s ChatGPT and Neo4j.