Available for opportunities

Goldius Leonard

AI Engineer based in Kuala Lumpur, Malaysia. I build intelligent systems that solve real business problems — from autonomous AI agents and large-scale fraud detection to computer vision pipelines processing millions of events.

Goldius Leonard
AI Engineer
LLM & RAG
Computer Vision
Agentic AI
Scroll
Goldius Leonard
About Me

Turning Data into Intelligence

I'm a data-driven AI Engineer with a strong background in computer science and a passion for solving complex problems using machine learning, NLP, and computer vision. With experience across the full AI lifecycle — from R&D to production deployment — I've led impactful projects in agentic AI orchestration, fraud detection, and large-scale social media monitoring.

My work has enhanced real-time decision-making systems, secured multimillion-ringgit investments, and built pipelines that process 50,000+ events daily. I hold a BSc in Artificial Intelligence, graduating Summa Cum Laude with a 3.92 GPA.

0
Years Exp.
0
Companies
3.92
GPA (Summa)
Kuala Lumpur, Malaysia
goldius.leonard@gmail.com
AI Engineer @ Deriv
BSc Computer Science (AI)
Download CV
Tech Stack

Skills & Expertise

A versatile toolkit spanning AI research, engineering, cloud infrastructure, and data systems.

Python
JavaScript
TypeScript
Java
C
PyTorch
TensorFlow
OpenCV
Scikit-learn
HuggingFace
OpenAI
LangChain
Google ADK
AutoGen
Gemini
Docker
Kubernetes
Git
Airflow
PostgreSQL
MySQL
Redis
ClickHouse
Supabase
AWS
GCP
Azure
Python
JavaScript
TypeScript
Java
C
PyTorch
TensorFlow
OpenCV
Scikit-learn
HuggingFace
Langfuse
OpenAI
LangChain
Google ADK
AutoGen
Gemini
Anthropic
Bedrock
LiteLLM
Docker
Kubernetes
Git
Airflow
Cloudflare
Okta
Metabase
FastAPI
PostgreSQL
MySQL
Redis
ClickHouse
Supabase
SQL Server
AWS
GCP
Azure
Career

Work Experience

Building impactful AI solutions across industries — from fintech fraud detection to national-scale content monitoring.

AI Engineer
May 2025 – Present
Deriv  ·  Cyberjaya, Malaysia
  • Spearheaded Bug Hunter Agent PoC using Microsoft AutoGen & Playwright for autonomous UI/UX anomaly detection, projected to reduce QA cycle times by 60%.
  • Engineered Partner Fraud Detector with ClickHouse & HDBSCAN, flagging 60+ fraudulent cases daily and preventing an estimated $50,000 in monthly revenue leakage.
  • Architected Partner Intelligence Hub secured via Cloudflare Zero Trust & Okta, reducing access management overhead by 40% through granular RBAC and App Attribute Access Control.
  • Built iCore Platform (Google ADK), deploying 14+ specialized AI agents via FastAPI/Kubernetes — cutting modification turnaround from 3 days to under 10 minutes and reducing support tickets by 80%.
AutoGen Google ADK ClickHouse Kubernetes Playwright Cloudflare ZT FastAPI
Pod Lead
Sep 2024 – Apr 2025
Userdata Sdn. Bhd.  ·  Kuala Lumpur, Malaysia
  • Achieved 60% improvement in chatbot response efficiency, reducing TTFC from 5 min to 2 min and raising accuracy satisfaction from 70% → 80% (DBS & AmBank use cases).
  • Scaled social media monitoring for MCMC & PDRM from 1,000 to 50,000 TikTok videos/day for national scam detection.
  • Led Azure-to-AWS migration reducing cloud costs by 25% while maintaining 100% service uptime with improved CI/CD.
AWS LLM CI/CD TikTok API Team Leadership
Data Scientist
Feb 2024 – Aug 2024
Userdata Sdn. Bhd.  ·  Kuala Lumpur, Malaysia
  • Developed AI chatbot using LLM & RAG, accelerating dashboard creation from months to 5 minutes, saving RM 100,000+ per dashboard.
  • Secured MYR 36 million in Series A funding and attracted potential MYR 300M in Series B by showcasing AI capabilities.
  • Built and optimized chart generation LLM agents supporting multiple database systems with a 70% satisfaction rate.
LLM RAG LangChain Python PostgreSQL
Junior Computer Vision Scientist
Feb 2023 – Jan 2024
Verihubs  ·  Jakarta, Indonesia
  • Improved ID card OCR accuracy from 85% to 95%, significantly enhancing document detection performance.
  • Increased document liveness detection accuracy to 90% under extreme lighting and occlusion conditions.
  • Reduced AI model pipeline latency by 30%, improving inference speed and client response times.
PyTorch OpenCV OCR FastAPI Python
Computer Vision Scientist Intern
Feb 2022 – Jan 2023
Verihubs  ·  Jakarta, Indonesia
  • Enhanced face liveness detection accuracy by 15% through research and dataset diversification.
  • Developed automated testing application reducing manual verification time by 40%.
  • Streamlined experiment documentation, reducing onboarding time for new engineers by 20%.
PyTorch TensorFlow Face Liveness Python
Academic

My Education

🎓
Bachelor of Computer Science (Artificial Intelligence)
Bina Nusantara University, Jakarta, Indonesia
Aug 2019 – Feb 2023
GPA: 3.92/4.00 — Summa Cum Laude
Best Graduate Award
Thesis: SSDG Framework with Feature Mining for Face Anti-Spoofing
Active in HIMTI organization (IT student body)
Projects

My Portfolio

A selection of ML, computer vision, and data projects from Kaggle, GitHub, and academic work.

News Classification
Machine Learning
News Classification & Translator
Multi-class text classification pipeline with automated language translation.
Email Spam Classifier
Machine Learning
Email Spam Classifier
NLP-based spam detection using classical ML algorithms with high precision.
Income Classification
Machine Learning
Income Classification
Census data analysis for income bracket classification using ensemble methods.
Exchange Rate
Deep Learning
USD to INR Exchange Rate Prediction
LSTM-based time series forecasting for currency exchange rate trends.
IMDB
Deep Learning
IMDB Movie Review Classification
Sentiment analysis using LSTM/Transformer networks on the IMDB dataset.
License Plate
Computer Vision
License Plate Detection
Real-time plate detection and recognition using YOLO and OCR techniques.
Face Anti-Spoofing
Computer Vision
Face Anti-Spoofing Detection
Advanced liveness detection based on the SSDG framework — thesis research.
Tweet Sentiment
Deep Learning
Coronavirus Tweet Sentiment
BERT-based sentiment analysis on COVID-19 Twitter dataset for public opinion mining.
Income Prediction
Machine Learning
Income Prediction
Regression-based income prediction using socioeconomic features from census data.
Sales Report
BI Dashboard
Sales Report – Power BI
Interactive customer sales analytics dashboard with drill-through and KPI tracking.
Amazing E-Book
Web
Amazing E-Book
Full-featured e-book platform website built with HTML, CSS, and JavaScript.
Ezy Foody
Mobile App
Ezy Foody
Food ordering mobile application developed for the Binus University project showcase.
Get In Touch

Let's Connect

I'm open to new opportunities, collaborations, or just a good conversation about AI. Feel free to reach out!