backend engineer — payment infrastructure
JamalKya Nanami
aka 菜々美
I build the routing layer between merchants and payment providers — the layer that isn't allowed to drop a transaction.
I work on the parts of a payment system that aren't allowed to fail quietly: transaction routing, retries, idempotency, audit trails. At Monarch Inti Teknologi I'm building the aggregator that lets one merchant integration reach six payment providers, with automatic failover when one goes down. Before that, at Firstudio, I shipped a citizen management system that now serves 24,000+ residents across departments.
experience
Jun 2026 — Present
Backend Developer
PT Monarch Inti Teknologi · West Jakarta, Indonesia
- Building a merchant aggregator engine — one integration point that reaches six Payment Junction Providers: Beyontech, Payara, Inacash, Moneyball, Teletrans, and Winpay.
- Designed the automatic routing logic that picks the best PJP per request, so payments keep flowing even when one provider goes down.
- Architected the system so onboarding a new PJP takes minimal code change — the business can add payment partners without reworking the codebase.
Jan — Jul 2025
Backend Developer
PT Firstudio · Semarang (Hybrid)
- Built a Citizen Management System handling 24,000+ resident records across 43 REST endpoints, so city admins manage population data across departments from one place.
- Secured the service with API key auth and rate limiting; shipped to a production VPS through Docker for repeatable releases.
Feb — Mar 2025
Web Developer Intern
PWM Jateng · Semarang
- Took two divisions' event registration, donation tracking, and public info pages fully online within two months.
Dec 2023
Student Mentor — Basic Programming
Universitas Dian Nuswantoro · Semarang
- Led a 3-day workshop teaching programming fundamentals to junior students.
education & research
2022 — Feb 2026
B.Sc. Informatics Engineering
Universitas Dian Nuswantoro (UDINUS) · Semarang
GPA 3.65 / 4.00 · 146 credits · 0 failing grades across 54 courses
December 2025
Best Paper Award
IEEE Xplore · ISEMANTIC 2025
Improving Customer Churn Prediction Using Domain-Driven Feature Engineering, Resampling, and CatBoost with Explainability Extensions
An end-to-end churn prediction pipeline on the IBM Telco Customer Churn dataset (7,043 samples, 26.54% churn): SMOTE for class imbalance, domain-specific engineered features, and CatBoost/XGBoost tuned with Optuna. The best model reached a 0.937 ROC-AUC and 0.863 F1-score. SHAP explains which features drive each prediction, DICE-ML generates counterfactual "what would change the outcome" explanations for retention teams, and a fairness audit surfaced a 22.55% bias toward senior citizens worth correcting for.
featured projects
Enterprise Banking Microservices System
"If a transfer succeeds but the notification service is down, the system retries automatically. No financial event is ever silently lost."
- Six independent services — transfers, auth (TOTP 2FA + JWT), accounts, audit logs, real-time notifications — each scales without affecting the others.
Payment Gateway Integration — Midtrans
"No duplicate charges. No data loss on crash. No balance conflicts."
- Detects double-submits before they charge twice, saves records locally before confirming so a crash never drops a transaction, and blocks two simultaneous payments on one account.
- Auto-pauses requests when Midtrans is unavailable instead of flooding the provider — keeps the system stable through outages.
stack
Languages
Frameworks & APIs
Databases & Messaging
DevOps & Architecture
available for new roles
Open to backend roles.
If you're hiring for payment infrastructure, distributed systems, or backend work in general — reach out directly.
© 2026 JamalKya Nanami (菜々美). All rights reserved.