
BEM Vokasi UNS
The official site for the faculty student executive board, covering news, events, and member management.
Software Engineer
I build the systems companies actually run on — e-procurement modules on .NET and Vue, SAP and Odoo ERP integrations, Web GIS and Power BI reporting — and the pipelines that ship them, from Jenkins and Argo CD on Kubernetes to automations in n8n. Alongside that I train computer vision models, design the interfaces in Figma, and lead the media team at BEM SV UNS. Informatics Engineering at Universitas Sebelas Maret.
Working with

Roles across procurement and ERP, product engineering, data, and design — several running side by side.
Building and maintaining e-procurement modules on a .NET backend with a Vue front end, and integrating procurement workflows with SAP to automate the business process end to end. Moved from a UI-focused brief into backend and enterprise database work.
Web-based geographic information systems for corporate clients: interactive maps and spatial data layers embedded in web applications, with responsive rendering and custom map queries built alongside the frontend and data teams.
Interactive Power BI dashboards tracking client performance metrics, turning complex datasets into visual reports that feed executive decisions.
High-fidelity prototypes and enterprise interface design for Cushman & Wakefield, including workflow mapping and wireframing, worked through with the development team to keep the designs buildable.
Leading front-end development on Solutionlabs' flagship products in React and Next.js, integrating against Go services and holding the line on code review and quality.
Interfaces for web and mobile, delivered as personas, wireframes, interactive prototypes and journey maps — then validated through usability testing and iterated on.
Digital campaigns for the company's software products, from market research and copy through to reading performance metrics back into spend.
Responsive, high-performance web interfaces, built with data engineering constraints in mind and with a media-production background feeding the visual side.
Visual assets for marketing and branding — social content, infographics and promotional material held to a consistent brand system.
Web applications end to end: scalable backend architecture, responsive interfaces, API integration and database query optimisation with cross-functional teams.
Technical support and troubleshooting across internal software platforms, keeping downtime down.
Operational and business data analysed into interactive Power BI dashboards, with KPIs and automated reports designed to take manual steps out of reporting.
Machine learning and vision models for object detection and classification on the university's national robot contest team, tuned for responsiveness under competition conditions.
Led the visual communication strategy for the faculty student executive board: 10+ media programmes, the official brand guidelines, and a team mentored through producing them.
Invited back to run an intensive session on media strategy, visual branding and creative workflow for the division.
Set the concept and curriculum for the webinar, guided the organising committee, and oversaw speaker and media-partner selection.
Ran a media team across three subdivisions — documentation, design and decoration — holding one visual identity across print, social and the venue itself.
Media partnerships and official communications for the festival, lifting press coverage by around 10%.
Grouped by the layer it belongs to — from enterprise systems and the delivery pipeline down to the microcontroller.
Procurement, SAP integration, and business systems
Services, data models, and the contracts between them
Product front ends, on the web and on the phone
Pipelines, clusters, and workflows that run themselves
Detection and classification models, trained and deployed
Dashboards, reporting, and web mapping
Microcontrollers and the firmware that drives them
Interfaces, prototypes, and visual identity
Computer vision at the intersection of embedded systems, healthcare, and cultural heritage.
A batik platform that identifies motifs from a photograph and synthesises new hybrid patterns by blending the aesthetic characteristics of two source motifs under an adjustable dominance weighting. Extends into virtual try-on and retrieval-augmented recommendation.
trisarabatik.my.idAn ESP32-CAM navigation system that detects obstacles with YOLOv8 and reports them through synthesised audio, combined with GPS positioning to support safe outdoor mobility for visually impaired users.
A convolutional neural network approach for classifying tympanic membrane sizes from otoscopy images, aimed at supporting earlier diagnosis of hearing disorders.
Enterprise and product software, the pipelines that deliver it, and the computer vision and embedded work alongside.

The official site for the faculty student executive board, covering news, events, and member management.

The earlier batik system: upload a photo and find matching patterns with a deep learning model, served from a Python backend behind a Next.js front end.


An ordering platform for a small florist business, covering the catalogue and order management.

A marketplace for second-hand goods with accounts, product listings, and purchasing between users.
A self-service queue system for a hospital: patients take a number at a kiosk, a public display calls it, and staff run the queue from a dashboard. Calls are broadcast live over WebSockets with Laravel Echo and Pusher, and tickets print straight from the browser. Laravel with Inertia and a React + TypeScript front end, with CI for lint, tests and deploy.
A clinic website with service pages, specialist profiles, articles and doctor appointment booking, built as a team. Nuxt with shadcn-style components, validated forms with Zod, and a Tailwind design system.
A membership app with tiers, points, vouchers and rewards — plus daily check-ins, missions, a customer survey, and QR codes both scanned and shown at the counter. Built in Flutter with animated transitions between screens.
A redesign of an open-source Laravel and React point-of-sale system. My part is the admin experience: a new dashboard layout, a sky-blue theme and navigation, a global AJAX loading state, and SweetAlert2 confirmations in front of destructive actions.
The invoice module of a procurement platform — a .NET service covering invoice creation, validation and its path through approval.
Court booking and scheduling, split across a Go API and a Next.js front end.
Stock and item tracking for a small operation, built on Laravel with a Vue front end.
A container delivery pipeline built end to end across three repositories. A Jenkins shared library runs every service through the same stages — checkout, a Trivy security scan of the source, tests, a BuildKit image build and push, a second Trivy scan of the image — then commits the new image tag to a GitOps repo and reports to Slack. Argo CD picks up that commit through an app-of-apps setup and syncs Kustomize overlays to Kubernetes, with autoscaling, disruption budgets, a security policy, and Gateway API routing. Proven on a Go service.

A platform for recognising, learning about and trying on Indonesian batik. It identifies a motif from a photo, blends two motifs into a new hybrid pattern with adjustable dominance, offers a virtual try-on and photobox, and serves recommendations through a RAG pipeline. Ships as an installable PWA with separate admin and partner dashboards.

YOLOv8 detection and localisation of brain tumors in MRI scans, supporting earlier and more consistent diagnosis.

YOLOv8 detection and counting of coconuts from field imagery, built to support agricultural monitoring and yield estimation.

Classification of ear conditions from otoscopic images, built to help healthcare professionals diagnose more quickly and consistently.

Computer vision detection of basketball hoops in images and video, built to support sports analytics and training applications.
Traffic anomaly detection over live CCTV streams. A custom-trained YOLOv8 accident model runs with ByteTrack tracking, and pixel positions are mapped to real-world distances to catch sudden stops and vehicles closing inside four metres. Events land in PostgreSQL, stream to the dashboard over WebSocket, and critical ones reach officers as WhatsApp alerts through Twilio.
Real-time hand tracking from a webcam feed, detecting landmarks frame by frame with OpenCV.

Smart glasses for visually impaired users. An ESP32-CAM runs YOLOv8 object detection, a GPS module handles positioning, and a DFPlayer speaks obstacles and directions aloud — built for safe outdoor mobility.

A Raspberry Pi otoscope that classifies the condition of the human tympanic membrane from captured images, distinguishing normal from abnormal membranes to assist early diagnosis of hearing disorders.
A smart-farming system for tomato growers. A MobileNetV3 model, quantised to TensorFlow Lite, classifies plant disease from a photo with a confidence score; a second model predicts harvest; and a dashboard tracks NPK, temperature, humidity and pH over MQTT. Includes a farming chatbot and QR-code login from the camera. FastAPI and MySQL behind a Next.js front end.
Serial-driven lighting control for Arduino, switching and dimming from a host script.
Open to internships, research collaboration, and freelance work.