Wilbert Chandra

WILBERTCHANDRA

Data Engineer & AI Engineer

Jakarta, Indonesia

ABOUT

I build robust data and AI systems.

I design scalable data engineering pipelines, distributed AI systems, and production-ready architectures that make data reliable and intelligent.

Binus University — GPA 3.87/4.00

PROJECTS

Selected Work
01

SatuDua

Emergency Response System

Compfest 2025 • React Native • FastAPI • Azure AI

Live calling with real-time transcription, AI summaries, and dispatcher dashboards. Built for the AI Innovation Challenge hosted by University of Indonesia.

View case study
02

UniPal

Conversational AI

PKM Indonesia • Python • Gemini • ElevenLabs

Speech-to-text → LLM → lifelike voice, built for low-latency interaction. Real-time conversational AI with seamless speech processing pipeline.

View case study
03

MRI Segmentation

Medical Imaging

Deep Learning • PyTorch • Computer Vision

Automated ROI segmentation with strong Dice/IoU on high-dimensional scans. End-to-end deep learning model for medical analysis.

View case study

EXPERIENCE

Career
Mar 2025 – Present

Data Engineer Intern

Samsung R&D Institute Indonesia

  • Led end-to-end internal AI service using RAG + LangChain.
  • Automated Airflow pipelines and built CI/CD with Docker + GitHub Actions.
  • Improved data governance and AWS architecture.
Jul 2024 – Sep 2024

Backend & AI Engineer

UniPal / PKM Indonesia

  • Architected backend service for real-time conversational AI interactions.
  • Built speech-processing pipeline with Google STT and ElevenLabs TTS.
  • Orchestrated core logic using Google Gemini for intelligent responses.
Jun 2024 – Dec 2024

Lead AI Engineer

MRI Segmentation Project

  • Spearheaded end-to-end deep learning model for medical scan segmentation.
  • Engineered CNN architecture for precise ROI identification.
  • Achieved high segmentation accuracy through rigorous hyperparameter tuning.

PUBLICATIONS

Research
Procedia Computer Science (Elsevier)

Multiclass Eye Disease Detection

Lightweight deep learning for 9 retinal conditions; EfficientViT achieved ROC-AUC 0.9780 with only 3 million parameters.

PyTorchEfficientViTComputer Vision
View DOI
IEEE Xplore

RSNA Degenerative Lumbar Spine Classification

Automated spinal condition detection using CNN ensembles on MRI datasets. Led technical implementation of multiple diverse architectures.

PyTorchMedical ImagingEnsemble Learning
View DOI

SKILLS

Tools & Tech

Languages

Python
SQL

AI & Frameworks

TensorFlow
PyTorch
LangChain
Airflow
VLLM
RAG

Cloud & Tools

AWS
Docker
GitHub Actions
Milvus

Skills

Data Modeling
Analysis
Visualization

Let's build
something.

Tell me what you're building. I'll reply with next steps and a clear plan.

Thanks.

Open to collaborations, research, and product work.