Andika Putra Perdana
Email
Open to Data & AI rolesJakarta, Indonesia

AndikaPutra Perdana.

I build machine learning systems that make it to production. Not just notebooks.

View workDownload CV

or email me at andikaputh1@gmail.com

99.34%
Best Model Accuracy
85 SUS
Usability Validation
9 Projects
Documented Projects
MLOps pipelineInteractive / 3D
  • ServingFastAPI · Docker
  • ModelXGBoost · TensorFlow
  • DataBatch · Stream
  • Python
  • PyTorch
  • TensorFlow
  • FastAPI
  • Docker
  • Kafka
  • PostgreSQL
  • MLflow
  • ChromaDB
  • Tableau
  • Next.js
  • SQL
01About

How I work

Background, technical focus, and how I turn ambiguous problems into systems people can actually use.

Portrait of Andika Putra Perdana
Andika Putra PerdanaData & AI

Jakarta, IndonesiaPortrait / 2026

“I enjoy the messy parts: dirty data, misleading metrics, and models that still have to be explained.”

7th-semester Informatics student at Universitas Gunadarma and a Data & AI Engineer. Experienced in public-sector data analysis at Diskominfo Depok, clinical ML research (SiPEKA, 99.34% accuracy), industry capstones (SignBank AI), and production systems spanning fraud decisioning, recommendation, and vector search.

I’m used to working end to end: cleaning raw data, choosing and evaluating models, then packaging them into APIs or dashboards that non-technical people can use. Leading capstone teams and supporting an HPC lab taught me to explain technical decisions without unnecessary jargon.

Education
B.Sc. Informatics, Universitas Gunadarma
7th semester · Active
Location
Jakarta, Indonesia
Open to remote / hybrid
Focus
Data & AI Engineering · MLOps · Analytics
Machine Learning, Data Engineering, Distributed Systems
Availability
Open to Data & AI roles
Response within 1-2 business days
02Skills

Toolkit

What I use day to day, grouped by function. Not just a wall of logos.

PRIMARY FOCUS

Data Engineering, Analytics & AI

Core Competencies

  • Machine Learning
  • Data Analysis & Visualization
  • Deep Learning
  • RAG & Vector Search
  • Statistical & Trend Analysis

Libraries & Frameworks

  • Pandas
  • NumPy
  • Scikit-Learn
  • PyTorch
  • FastAPI
  • Flask

Visualization & Tools

  • Tableau
  • Jupyter
  • Docker
  • Git
  • Canva

SYSTEMS & PIPELINES

Backend & Data Pipelines

Languages & Core

  • Python
  • SQL

Frameworks & API

  • FastAPI
  • Streamlit
  • RESTful API

Data Pipelines & Scraping

  • Redpanda / Kafka
  • Selenium Scraping
  • SQLite
03Experience

Track record

Technical, research, and leadership roles that shaped how I work.

  1. Sep 2026 - Present

    Depok, West Java · On-site

    Ongoing

    Data Analyst Intern

    Data Analyst InternDepartment of Communication and Information Technology (Diskominfo), Depok City

    • Contributed to the management, processing, and analysis of sectoral statistical data for the City of Depok to support effective public information delivery.
    • Built interactive Tableau dashboards to visualize key statistical indicators and communicate data-driven insights.
    • Designed data-driven infographics in Canva to present statistical information in a clear, communicative, and visually engaging way.
    • Transformed raw data into informative visual outputs to support public information dissemination and data-driven decision-making.
    • Data Analysis
    • Tableau
    • Data Processing
    • Canva
    • Sectoral Statistics
    • Data Visualization

  2. Oct 2026 - Present

    Remote

    Ongoing

    Data Science & AI Capstone Team Lead

    Team Leader & AI EngineerAsah Bootcamp by Dicoding

    • Led a cross-functional team in designing and building data analytics and AI-based systems for local MSMEs.
    • Team Leadership
    • Data Analytics
    • AI System Design

  3. Feb 2026 - Jul 2026

    Remote

    Cohort: AI Engineer

    AI & Deep Learning ParticipantCoding Camp powered by DBS Foundation

    • Studied Machine Learning, Deep Learning, NLP, Computer Vision, and MLOps using TensorFlow.
    • Built SignBank as the capstone project: a real-time Computer Vision sign-language-to-text platform that helps bank tellers communicate with Deaf customers.
    • TensorFlow
    • Deep Learning
    • Computer Vision
    • NLP
    • MLOps
    • Python

  4. Oct 2024 - May 2025

    Depok, West Java · Hybrid

    Assistant of High Performance Computing (HPC)

    Teaching Assistant & Lab SupportHPC Universitas Gunadarma

    • Provided technical support, supercomputer/DGX access troubleshooting, and lab guidance for students.
    • Coordinated lab operations, focus group discussions (FGD), and divisional events.
    • High-Performance Computing
    • Python
    • Data Science
    • Troubleshooting
04Work

Selected work

Nine systems built and documented: from clinical ML and fraud decisioning to mission-aware recommendations.

UNDERGRADUATE THESISFlagship
CLINICAL ML · XGBOOST

Academic & Machine Learning

01

SiPEKA: Early Preeclampsia Risk Prediction

Undergraduate Thesis: Universitas Gunadarma

An XGBoost + SMOTE preeclampsia risk prediction web app with 99.34% accuracy and 0.9993 AUC. Includes clinical safety-override rules, automatic MAP/BMI calculation, and print-ready medical reports. Validated with an SUS score of 85 (Excellent).

Accuracy: 99.34% | SUS: 85 (Excellent)

  • Python
  • Flask
  • XGBoost
  • SMOTE
  • Scikit-Learn
CodeLive demo
DBS CAMP CAPSTONEFlagship
COMPUTER VISION · REAL-TIME

Bootcamp Capstone & AI

02

SignBank: AI Sign Language Translation

Capstone Project: Coding Camp by DBS Foundation

A real-time Computer Vision and Deep Learning sign-language-to-text platform that helps bank tellers communicate with Deaf customers.

Real-time Vision Inference

  • Python
  • TensorFlow
  • Deep Learning
  • Computer Vision
  • OpenCV
  • Next.js
Private repositoryLive demo
MISSION-AWARE RECOMMENDERFlagship
HYBRID RANKING · ILP

MLOps & AI Engineering

03

RASA: Relevance-Aware Shopping Assistant

Live demo · synthetic retail data

A mission-aware product recommendation system that combines content, collaborative, and mission signals, then optimizes a basket against budget, category, inventory, and merchant constraints.

97.5% budget feasibility · 100% constraint coverage · 78 ms p95

  • Python
  • FastAPI
  • React
  • TF-IDF
  • ILP Optimization
  • Docker
CodeLive demo
REAL-TIME FRAUD OPSFlagship
FRAUD DECISIONING · SSE

MLOps & AI Engineering

04

Sentinel: Real-Time Fraud Engine

Synthetic demo for Meridian Pay

A real-time fraud decisioning engine with Isolation Forest anomaly scoring, 15 explainable policy rules, alert triage, and a live operations console.

ROC-AUC 0.9801 · 15 policy rules · live SSE feed

  • Python
  • FastAPI
  • Isolation Forest
  • React
  • SQLite
  • Docker
  • Render
CodeLive demo
RAG & VECTOR SEARCH
RETRIEVAL · VECTOR SEARCH

GenAI & LLM Workflows

05

AI-Powered PDF Q&A (RAG Lite)

An interactive PDF question-answering app using Hugging Face local embeddings, ChromaDB vector search, and the Groq LLM inference API.

Local Embeddings + Fast LLM

  • Python
  • Streamlit
  • LangChain
  • ChromaDB
  • Hugging Face
  • Groq API
Code
NLP MICROSERVICE
NLP · REST API

ML Microservices & NLP

06

Hugging Face FastAPI Sentiment Microservice

A high-performance REST API microservice for text sentiment analysis using a pretrained Hugging Face Transformer and Pydantic validation.

Swagger UI + Local Inference

  • Python
  • FastAPI
  • Transformers
  • PyTorch
  • Pydantic
  • Uvicorn
Code
MACHINE LEARNING
CLUSTERING · PCA

Machine Learning & Analytics

07

Credit Card Customer Segmentation

A customer profile segmentation app using K-Means and PCA, deployed to Streamlit Community Cloud.

Silhouette Score: 0.74

  • Scikit-Learn
  • Streamlit
  • Pandas
  • Python
CodeLive demo
AI DATA PREPROCESSING
DOCUMENT → MARKDOWN

Tools & AI Preprocessing

08

MarkItDown Web Converter (Doc2MD)

A privacy-first web converter that turns PDF, DOCX, PPTX, and XLSX files into clean Markdown ready for LLM or RAG pipelines.

In-Memory Batch Processing

  • Python
  • FastAPI
  • Microsoft MarkItDown
  • React
  • Vite
CodeLive demo
WEB APP & ANALYTICS
WEB APP · TYPING ENGINE

Interactive Web Apps

09

TypeHybrid: Touch Typing Platform

An interactive touch-typing practice app with a learning curriculum, visual keyboard guide, real-time statistics, and mechanical keyboard sound effects.

Adaptive UI + Audio Profiles

  • Next.js
  • Tailwind CSS
  • Web Audio API
  • Vercel
CodeLive demo
05Contact

Have a problem worth solving?

Share the context and the goal. I’ll respond with a realistic approach, even if the smartest solution isn’t machine learning.

Email

andikaputh1@gmail.com
Availability

Open to Data & AI roles

Full-time · contract · research collaboration

Location

Jakarta, Indonesia

Remote / hybrid · WIB (UTC+7)

Response

1-2 business days

Via email, LinkedIn, or the form

Education

B.Sc. Informatics, Universitas Gunadarma

7th semester · Active

Elsewhere

Download CV (PDF)