Johannes Baptista Adiatmaja Pambudi

About

Data Scientist based in Indonesia, building end-to-end ML solutions from data pipelines to production models.

Education B.S. Informatics, UKDW GPA 3.75/4.00 — 2024

Certification

TensorFlow Developer Google — 2023

Skills

Python

SQL

MySQL

Java

PHP

PyTorch

TensorFlow

Keras

scikit-learn

Pandas

NumPy

Streamlit

ChromaDB

Docker

Git

n8n

Python

SQL

MySQL

Java

PHP

PyTorch

TensorFlow

Keras

scikit-learn

Pandas

NumPy

Streamlit

ChromaDB

Docker

Git

n8n

Experience

Data Scientist

Sigma Solusi Indonesia

Aug 2024 — Present

Architected Dockerized RAG chatbot (ChromaDB, sentence-transformers) for B2B FMCG platform — semantic product search and 35+ intent classification types
Fine-tuned a 9B model for WhatsApp deployment via knowledge distillation (Llama-3.1 70B → Gemma-2 9B) with LoRA
Built MLOps retraining pipeline via n8n — 10+ iterations with A/B validation across 154 structured test scenarios
Developed demand forecasting pipelines (PyTorch, SQL) and BI solutions including warehouse simulations and stock balancing

Machine Learning Engineer

Beehive Drones Intern

Aug 2022 — Jan 2023

Developed data pipeline for aerial imagery processing for carbon value estimation
Engineered deep learning model achieving 93% improvement over baseline

Projects

2026

LLM Distillation Pipeline

Indonesian-Javanese Language Model Fine-tuning

Distilled a 70B teacher model into a 9B student model via a three-zone Docker pipeline on a restricted GPU server with no Python runtime.

Knowledge Distillation LoRA Unsloth Docker

2025

ChromaDB RAG Chatbot

B2B FMCG Semantic Search & Intent Classification

Architected a Dockerized RAG system using ChromaDB and sentence-transformers for a B2B platform — enabling semantic product search across colloquial Indonesian names and 35+ intent classification types.

RAG ChromaDB sentence-transformers Docker

2025

FMCG Demand Forecast

Indonesian FMCG Inventory Planning Pipeline

Dual-pipeline deep learning system producing probabilistic P10/P50/P90 demand forecasts via an LSTM encoder + autoregressive GRU decoder with additive attention, Indonesian FMCG calendar features, and a purchase recommendation engine.

Time-Series Forecasting PyTorch Quantile Regression STL Decomposition

2024

Gender Classification

Fingerprint Analysis — Undergraduate Thesis

Compared three Gabor filter configurations against raw fingerprints using a custom CNN and 5-fold cross-validation on two datasets, achieving up to 73.46% weighted F1 with the default parameter set.

CNN Gabor Filters TensorFlow K-Fold CV

2023

ASAH

Waste Classification Mobile App

Achieved 94% accuracy using MobileNetV2, optimized for real-time inference on Android.

MobileNetV2 TensorFlow Android