Sign in Subscribe

You Can Now "Interview" Me: Introducing My AI Digital Twin

You Can Now "Interview" Me: Introducing My AI Digital Twin
Photo by Theo / Unsplash

I am back to blogging after a long time!

If you've looked at the navigation bar on my blog, you might have noticed something nestled: Chat with me.

As much as I’d love to jump on a call with everyone who visits my portfolio to discuss data science, AI, or career paths, physics unfortunately limits me to one place at a time.

So, I decided to build the next best thing: an AI "Digital Twin." It’s a RAG-based chatbot capable of answering questions about my background, skills, and experience, 24/7.

Here is a quick look at why I built it and the tech stack powering it "under the hood."

The Goal: A Grounded Persona

I didn’t want a generic ChatGPT wrapper that hallucinates answers about my life. The goal was to create an agent that acts authentically as "Zin Ko Hlaing."

To achieve this, the bot uses RAG (Retrieval-Augmented Generation). Before it answers a user's question, it first looks up relevant information from a curated knowledge base.

The primary source of truth for this bot is my current LinkedIn profile (exported to PDF) and a structured summary text file of my career highlights. If you ask about my experience with a specific ML stack, it’s pulling that answer directly from my verified history, not general internet knowledge.

Furthermore, if you ask me a question that the bot cannot answer using information from my LinkedIn profile, it will immediately ping me on my phone using Pushover app.

Unknown question that the bot cannot answer
And it is pushed immediately to my phone!

Under the Hood: The Tech Stack

Getting this running live involved stitching together a few key services:

  • The Brain: OpenAI’s gpt-4o-mini model.
  • The App: Built with Gradio and Python.
  • The Hosting: The app runs as a persistent "Space" on Hugging Face.
  • The Integration: Embedded into this Ghost blog via an Iframe.
  • The Alerts: Pushover for real-time notifications on my phone.

The Data Flow Diagram

Here is how the logic flows from your question to my phone:

Give It a Try

It’s live right now. Head over to the Chat with me page and ask it about my experience with Data Science or what projects I’ve worked on.

And go ahead and try to stump it. If you do, my phone will let me know!