Blog

Meet Virtual Mark

Written by Mark Tebault | Mar 2, 2026 2:04:17 AM

 

Introducing Virtual Mark — My AI Interview Agent

As AI becomes central to modern product and technology strategy, senior leaders are increasingly expected to bring practical, real‑world understanding of AI/ML, RAG, and enterprise agent architectures.

To demonstrate my depth in this space, I built Virtual Mark — an AI interview agent powered by Retrieval‑Augmented Generation (RAG), trained on my full career history, accomplishments, and behavioral responses.  Click the button below to launch Virtual Mark.

If you’re a hiring manager or recruiter looking for a leader who can bridge strategy, execution, and modern AI capability, I invite you to test it out. It’s a small example of how I think, architect, and deliver.

Anyone can say they understand AI.
I wanted to prove it by building a working solution that reflects the skills senior technical leaders need today.

I designed the system, shaped the data approach, and guided the architecture — the same leadership mindset, clarity, and technical literacy required to drive AI initiatives at the Director level and above.

For the technical folks:

https://github.com/TebaultTechnologyGroup/virtual-mark

Frontend
  • React 19 with TypeScript
  • Vite for bundling
  • Hosted as a Static Web App on Azure
Backend
  • Python 3.11 Azure Function (v2 programming model)
  • azure-ai-projects 2.0.0b1 SDK
  • openai 1.x client library
  • azure-identity for Service Principal authentication
AI / Agent
  •  Azure AI Foundry agent with RAG
    • gpt-4o
    • text-embedding-3-small
  • OpenAI Responses API
  • Model hosted in Azure East US 2
Authentication
  • Azure Entra ID Service Principal
  • Role: Azure AI Developer assigned at both the Foundry account and project scope
Deployment
  • GitHub Actions CI/CD
  • Azure Static Web Apps (Free tier)
  • Frontend built by the workflow, Python API deployed alongside it
  • Custom domain via IONOS CNAME → Azure Static Web Apps