Profile
Ang Kai Liang
Building Actively

Ang Kai Liang

AI Engineer

Get In Touch
Dashboard

Key Metrics

LocationKuala Lumpur, MY
EducationAsia Pacific University
Current FocusLocal-first AI Agents
Public Projects8+ Repos

About Me

I'm an AI engineer who likes turning ambitious ideas into tools people can actually use. My work sits between machine learning, agentic systems, computer vision, and product engineering, with a strong focus on making technical systems feel clear, reliable, and practical. My current research and development focus is Monaw, a local-first Windows AI agent that works inside the user's machine with controlled access to files, shell commands, browser automation, memory, schedules, and MCP tools. Building it has sharpened how I think about permissions, orchestration, tool use, and dependable desktop automation. I have also built applied AI products across different domains, including Budgie for personal finance, ThinkStorm for AI-assisted ideation, medical imaging models for MRI and X-ray analysis, and real-time computer vision workflows. Those projects taught me to connect model performance with product experience, not treat them as separate concerns. Outside engineering, photography keeps me attentive to composition, timing, and detail. That creative habit carries into how I design interfaces and shape the feel of the products I build.

#AgenticAI#ComputerVision#Flutter#ProductEngineering
CURRENT FOCUS

From Models to Products

Agentic AI & Automation

Building Monaw as a local-first desktop agent with tool use, memory, permissions, automation, and practical workflow control.

Applied Vision Systems

Fine-tuning and shipping vision workflows for MRI, X-ray, license plate recognition, and geospatial analysis.

Tech Stack

Python
PythonCompetent
TypeScript
TypeScriptCompetent
React
ReactCompetent
Next.js
Next.jsCompetent
FastAPI
FastAPICompetent
Flutter
FlutterCompetent
Dart
DartCompetent
Agentic AICompetent
Google Gen AI
Google Gen AICompetent
Machine Learning
Machine LearningCompetent
Deep Learning
Deep LearningCompetent
Computer Vision
Computer VisionCompetent
PyTorch
PyTorchCompetent
Gradio
GradioCompetent
Data Analysis
Data AnalysisCompetent
Jupyter Notebook
Jupyter NotebookCompetent
Docker
DockerCompetent
Git
GitCompetent

Engineering Philosophy

Modular & Clean

I prefer systems that are easy to reason about: clear modules, explicit boundaries, and implementation details that future me can still understand.

AI With Control

I like intelligent systems that stay useful and accountable, especially when agents handle files, tools, memory, automation, and user permissions.

Product Feel Matters

A strong model still needs a thoughtful interface. I care about performance, motion, layout, and small details that make technical products approachable.

More About Me

View my journey