Riantama Putra, Information Systems & Technology @ ITB

Data science and AI engineer in Bandung. I build intelligent systems from model design and validation to production-ready software.

EngineeringIntelligence

Profile

About

Evidence-first systems, built to ship.

I work across machine learning, data science, and full-stack engineering. The throughline is measurable evidence: honest validation, competition results, and software that turns a model into something people can use.

Forecast analysis dashboard showing model comparisons and seven-day demand projections.

Validate forward, not backward.

Forward-chaining evaluation and leak-free splits keep forecasting and classification results honest.

Blue-lit computer processor mounted on a circuit board.

Make models earn their place.

Architecture choices are tied to leaderboard scores, error metrics, and clear attribution rather than novelty alone.

Infrastructure health dashboard beside server racks and a cloud diagram.

Ship beyond the notebook.

APIs, authentication, databases, and deployment turn experiments into products that can survive real use.

Presenter leading a technical workshop for participants using laptops.

Lead the work around the work.

Technical delivery sits alongside sponsorship, creative, public-relations, mentoring, and field leadership at ITB.

Engineering range

Domain

Applied AI and Forecasting

Computer vision, graph attention, gradient boosting, and feature engineering for noisy real-world prediction problems.

Face Anti-Spoofing, Bengawan Solo, City Traffic

Decision Systems

Learning-to-rank, cross-encoder reranking, risk classification, and interpretable regression built around careful validation.

Wikipedia Next-Click, Phishing Detection, Soil Regression

Fullstack Products

Role-based applications, REST APIs, AI-assisted workflows, and production software across web, desktop, and CLI environments.

Autopilot, OceanEarn, BelajarDuluDek

Python
TypeScript
JavaScript
C
Java
SQL
PyTorch
LightGBM
XGBoost
CatBoost
Transformers
timm
SHAP
YOLO
scikit-learn
Pandas
NumPy
React
Next.js
Tailwind CSS
Framer Motion
Django
FastAPI
PostgreSQL
Prisma ORM
Supabase
Docker
Git
Python
TypeScript
JavaScript
C
Java
SQL
PyTorch
LightGBM
XGBoost
CatBoost
Transformers
timm
SHAP
YOLO
scikit-learn
Pandas
NumPy
React
Next.js
Tailwind CSS
Framer Motion
Django
FastAPI
PostgreSQL
Prisma ORM
Supabase
Docker
Git

Selected projects

Works

Featured work

05 projects

Autopilot Asia Hackathon page with Kuala Lumpur finals details.

AI & Backend Engineer

HR Retention Command Center

A multi-operator AI employee coordinating a 90-day onboarding journey, detecting attrition signals, and routing confidential escalations to HR, IT, and Payroll through Slack.

Slack API
Supabase
Typeform

Finalist — Autopilot Asia Hackathon 2026

Machine Learning Engineer

Bengawan Solo Forecasting

Forecasted 242-day water levels across 30 river monitoring posts with a graph-attention hydro embedding encoder fused with LightGBM and strict forward-chaining validation.

Python
PyTorch
LightGBM
GAT

1.41971 private RMSE · Finalist — SSDS 2026

Computer Vision Engineer

Face Anti-Spoofing

A dual-stream DINOv2 and ConvNeXtV2 system with Laplacian high-frequency enhancement and asymmetric gated fusion for limited-data spoof classification.

Python
PyTorch
DINOv2
ConvNeXtV2

0.982 private leaderboard score · Finalist — DAC FindIT! 2026

WaProject image placeholder

Computer Vision Engineer

Waste Material Classification

Classified 26,527 waste images into recyclable, electronic, and organic categories with a frozen DINOv3 ViT-7B backbone and lightweight multi-branch head.

Python
PyTorch
DINOv3
timm

99.8869 Macro F1 · Rank 24 of 544 — BDC Satria Data 2026

Fullstack Developer

BelajarDuluDek

A production learning platform for Indonesian high-school students with AI chat, flashcard generation, quizzes, tryouts, and JWT authentication.

Next.js 16
TypeScript
Prisma
PostgreSQL

Deployed to production

More work

07 projects

Data Scientist

Traffic Speed Forecasting

Forecasted three horizons across 1,260 road segments using residual LightGBM ensembles per regime over a segment-level mean-reversion anchor.

38.831 private MSE · Semifinalist — Datathon 2026

Python
LightGBM
Feature Engineering

Machine Learning Engineer

Wikipedia Next-Click

Reconstructed an internal link graph with OCR, then reranked candidate links using LightGBM learning-to-rank and cross-encoder semantic scoring.

0.448 private accuracy · Semifinalist — Datathon 2026

Python
LightGBM
OCR
Sentence-Transformers

Data Scientist

Phishing URL Detection

A leak-free classifier built from intra-URL lexical features after identifying duplicate-driven leakage across 47.6% of the source rows.

0.899 F1 · 0.966 ROC AUC · FIT Competition 2026, Track III

LightGBM
XGBoost
CatBoost
SHAP

Data Scientist

Soil Organic Regression

An OOF-validated regression pipeline combining target transforms, hierarchical target encoding, and a greedy forward-selected ensemble.

10.972 private RMSE · DSA COMPFEST 18

Python
LightGBM
XGBoost
CatBoost

ML Systems Engineer

Open Transcript

A fully offline transcription CLI with speaker diarization, CPU-oriented int8 inference, resumable batch processing, and TXT, JSON, SRT, and VTT exports.

Python
faster-whisper
pyannote.audio
CTranslate2

Fullstack Developer

OceanEarn

A maritime waste reward platform with three auth roles, Google SSO, YOLO waste estimation, duplicate detection, and nearest collection-point lookup.

19 REST API endpoints · I/O Festival 2026

React
Django
PostgreSQL
YOLO

Authentication Developer

DearDiary

A desktop system for internal innovation projects, including idea tracking, experiment logs, prototype history, and role-based authentication.

Java
JavaFX
PostgreSQL
Maven

Operating history

Experience

Feb. 2025 – Apr. 2025

Backend Developer

Wisuda April ITB 2025

Developed the backend for an anonymous message-submission feature with Next.js, TypeScript, and Prisma ORM on the T3 stack.
Built the Menfess service as a standalone module ready for integration into the main user profile.

Aug. 2024 – Oct. 2028

B.Sc. Information Systems & Technology

Institut Teknologi Bandung

Current cumulative grade point average: 3.64 out of 4.00.
Coursework and projects span machine learning, software systems, data engineering, and full-stack development.

2024 – Present

Student Leadership

Institut Teknologi Bandung

Head of Sponsorship for Arkavidia 11.0 and Head of Creative for Aksi Angkatan SPARTA HMIF ITB 2025.
Mentor and Field Committee for OSKM ITB 2025; led public relations for Formatur STEI-K 2024 and documentation for TPB Cup STEI 2024.

Build something real

Contact

For applied AI and fullstack systems.

Reach me for internships, technical collaboration, or project conversations. My source, profile, and inbox are one step away.

GitHub LinkedIn Email me