Thadoe Hein / Data & AI Engineer

I build with data and AI.

I build AI systems for real-world use, from enterprise platforms and document processing to RAG, evaluation, and data quality.

Explore my experience Banking · products · research

How I work / 01

Turning data and AI into useful systems.

Organize

Useful data

Build

AI + context

Deliver

Useful system

ClearEvaluatedDeployed
Data NLP RAG Data Governance Computer Vision AI Systems Evaluation

Professional experience / 2024—now

Where I've put AI to work.

Across banking, products, teaching, and research, I've worked on the systems and details that make AI useful in practice.

  1. Jun 2026 - Current

    AI Engineer

    CB Bank PCL · Data and AI Team

    Contributed to an enterprise AI platform and AI-assisted document processing, helped define data-quality standards, and coordinated technical work with external partners.

    Enterprise AI · Banking

  2. Jan—Jun 2026

    Freelance Software Engineer

    SoeMindAI, Inc.

    Built three AI products across education and recruitment, including study tools, presentation automation, and job preparation workflows.

    3 products · ~30% lower cost

  3. Jan—Jun 2026

    AI & Machine Learning Instructor

    Kiddee Lab

    Taught students aged 10–15, built reusable Python ML labs, and researched a vision pipeline for digitizing physical student records.

    Research accepted at iSTEM-Ed 2026

  4. Jun—Aug 2025

    Machine Learning Engineer Intern

    Language Understanding Laboratory

    Led a pilot to fine-tune mT5-base on a Burmese question-answering dataset and built its training and evaluation pipeline from scratch.

    7k train · 1.7k evaluation examples

  5. Jan 2024—Oct 2025

    AI Developer

    Intelligent Systems Laboratory

    Built NLP and RAG pipelines for low-resource languages, designed chatbot architecture, and mentored students through development and deployment.

    Research · mentoring · delivery

What I work with / 06 areas

How I build.

These are the areas I keep coming back to. Open a card to see the tools I use and a real example from my work.

Selected projects / 2025

A few things I've built.

Each one started with a problem I wanted to understand. Here is what I tried, how the pieces fit together, and what I learned.

01Visual retrieval

Helping people find themselves in hours of event footage.

Upload a selfie, search indexed event videos, and jump to the moments where that person appears.

ArcFaceQdrantAWSNext.jsPostgreSQL

90%

less processing time

02Knowledge systems

A document assistant built around retrieval, not guessing.

A conversational document tool that parses a file, retrieves relevant passages, and carries context across follow-up questions.

Google PaLMChromaDBFirebaseStreamlitPython

RAG

retrieval to feedback

03Document vision

Training receipt detection without labeling everything by hand.

A receipt-text localization pipeline that used Azure AI Vision to bootstrap annotations before fine-tuning YOLOv11.

YOLOv11Azure AI VisionPythonOpenCV

.93

detection accuracy

Something I've learned

“I enjoy the part between a promising model and a system someone can rely on.”