Hi, my name is
Dinie Muzaffar.
I turn data into reliable pipelines, models, dashboards, and insights.
I'm a Data Developer & Oracle PL/SQL Developer based in Kuala Lumpur with 2+ years across ETL, data engineering, analytics, and machine learning. I build end-to-end PL/SQL pipelines for Claims, AML, and E-Invoice, and I'm an aspiring Data Engineer / Data Scientist growing a public portfolio in Python and the modern data stack.
About Me
Hi! I'm Dinie, a data & PL/SQL developer at Maybank Ageas Holding Berhad (MAHB), Malaysia's banking & insurance GLC. For the last 2+ years I've been building end-to-end Oracle PL/SQL ETL pipelines for Claims, AML, and E-Invoice modules, and turning messy financial data into BI dashboards and data-driven decisions. My background is a Computer Science degree from UiTM and an APU-backed data science programme.
I'm passionate about working with data — analysis, engineering, or science, the domain doesn't matter. From CNN/LSTM academic projects to production claims, AML, and LHDN E-Invoice pipelines, what drives me is the same thing wherever I go: clean, honest data people can actually rely on.
Right now I'm levelling up in Power Platform (Power BI, Power Apps, Power Automate), the modern data stack (dbt, DuckDB, Airflow, Spark), and Java / Spring Boot & Docker, and building a public portfolio to show it works. Here's what I use daily, plus what I'm currently learning:
Core — used daily & defensible
- Oracle PL/SQL & Oracle 19c
- SQL & SQLite & Supabase
- Python (Pandas, NumPy)
- scikit-learn & TensorFlow
- ETL & Data Modelling
- Data Reconciliation
- Streamlit, Altair & FastAPI
- Power BI & Excel
- AML / BNM / LHDN
- CNN/LSTM (Deep Learning)
Currently learning
- Power Apps
- Power Automate
- DAX & Power Query (M)
- Java & Spring Boot
- Docker & Containerisation
- Modern data stack (dbt, DuckDB, Airflow, Spark)
📍 Kuala Lumpur, Malaysia
Where I've Worked
Data Developer / PL/SQL Developer @ Maybank Ageas Holding Berhad (MAHB)
June 2024 - Present · Banking & Insurance (GLC) · Kuala Lumpur
- Develop and maintain Oracle PL/SQL packages, stored procedures, functions, and triggers supporting Claims, AML compliance, and LHDN E-Invoice financial workflows.
- Automate the Regular Watchlist process by implementing PL/SQL triggers, procedures, packages, and batch jobs, working with the ETL team to handle ~3 million records daily, replacing a manual process previously managed by L2 Support.
- Design and implement ETL and batch-processing scripts to extract, validate, transform, and load financial data across Oracle modules.
- Delivered a 90% improvement in data accuracy during AML data remediation through profiling, cleansing, reconciliation, and restructuring of inconsistent records.
- Develop and iteratively enhance complex SQL for BNM regulatory data extraction, refining the data model across versions to meet evolving business requirements.
- Maintain the AML watchlist database, inserting and managing UN Security Council sanctions records for PEP, PNG, and fraud-prevention screening.
- Support LHDN E-Invoice integration and perform query optimisation & PL/SQL tuning in Oracle Database 19c.
- Conduct UAT testing and data-quality validation for regulatory reporting extracts (policy holder, beneficiary, product info); document the Claims Lifecycle System (schemas, batch processes, troubleshooting).
- Use AI coding tools (GitHub Copilot, Claude) for code review, SQL optimisation, and documentation.
Things I've Built
Featured Projects
Featured Project — F1 Data
Box-Box Analytics: F1 Race Pace Dashboard
An interactive Formula 1 dashboard for analysing race pace, tyre degradation, and pit-stop strategy using official FIA timing data. Visualise lap-time evolution, compare drivers head-to-head, and inspect tyre compounds & pit-stop timelines. ETL pipeline loads 42,300+ lap records from 38 races across the 2025-2026 seasons into a Postgres warehouse (Supabase); a companion Power BI dashboard reads from the same warehouse via ODBC.
- Python
- FastF1
- Pandas
- PostgreSQL
- Power BI
- Streamlit
Featured Project — Public Transit
LRT Kelana Jaya Monitor
A deployed insights dashboard for the Kelana Jaya LRT line combining official data.gov.my ridership (2019-2026) with 800+ real public posts from Threads, classified into crowding, delay, disruption, and other signal categories via keyword-based classification. Built privacy-first: the deployed dashboard reads only aggregate statistics — never raw post text or author identities.
- Python
- Streamlit
- Altair
- FastAPI
- SQLite
- Supabase
- pytest
Featured Project — Personal Engineering
Reviving My Old Samsung Galaxy S10+
A documented case study: an old S10+ stuck at 1% battery with "Charging stopped — temperature too low". Used ADB to find a false -20°C reading, tested a temporary software override, then did the real repair — disassembly, a replacement part, reassembly, and a revived phone.
- ADB
- dumpsys
- PowerShell
- Hardware
- Documentation
Featured Project — Computer Vision
Meko & Lily Classifier
A deployed Streamlit app that tells two family cats apart from any other cat. Trains a two-stage fine-tuned MobileNetV2 on 5,166 privacy-stripped images with macro-recall early stopping and occlusion augmentation — val accuracy 0.994 / macro recall 0.984 — plus YOLOv8 detection and snapshot & live-webcam (WebRTC) modes. Fixed a repeated "Lily tagged as Meko" mislabel, lifting holdout passes from 3/7 to 5/7.
- Python
- TensorFlow
- MobileNetV2
- YOLOv8
- OpenCV
- Streamlit
Lip-Reading Prediction System FYP
Final-year deep-learning project using CNN/LSTM to predict speech from lip movement; reached high test accuracy while keeping training runs reasonably short.
- Python
- TensorFlow
- CNN
- LSTM
Student Enrollment Analytics
Data-driven insights into university enrolment trends in Malaysia (2016–2020) for strategic planning and policy-making.
- Python
- Data Analysis
- Power BI
New project — in the works
Pieces are being assembled in the lab. The full reveal is coming soon.
● In BuildEducation & Certifications
B.Sc. Computer Science (Hons.)
Universiti Teknologi MARA (UiTM) Shah Alam
March 2022 - June 2024
- Final Year Project: Designed a lip-reading prediction system using CNN/LSTM; it reached high accuracy on test data without long training runs.
- Built Power BI dashboards from cleaned, combined datasets to surface business insights.
- Applied RapidMiner for data preprocessing, transformation, and pattern extraction from raw datasets.
Diploma in Computer Science
Universiti Teknologi MARA (UiTM) Cawangan Terengganu
June 2019 - February 2022
APU Certified Data Science Programmer (ACDSP)
Asia Pacific University of Technology & Innovation (APU)
December 2023 - June 2024
Fully funded by Yayasan Peneraju Pendidikan Bumiputera (SPRINT programme).
Licenses & Certifications
Get In Touch
I'm currently open to data analytics, data science, data engineering, business intelligence, ETL, and Oracle database roles.
Whether you want to collaborate on a data project, discuss an opportunity, or just say hi, my inbox is always open. I'll get back to you as soon as I can.