Physics graduate who ended up really liking code. Currently focused on data analytics, machine learning, and contributing to open source.
I studied Instrumentation Physics at UNNES (GPA 3.59) and spent most of my time building things that involved sensors, data, and eventually machine learning.
At BRIN I worked on a computer vision project using YOLO11 to detect fungal infections on leaves. We got to 89.6% precision, which I'm pretty happy with. Before that, I did a virtual internship at Kimia Farma where I built sales analytics dashboards using BigQuery and Looker Studio.
These days I'm also writing C# and contributing to open source .NET projects. It started as something new to learn and now I actually enjoy it.
- 📍 Bekasi, Indonesia
- 🎓 Fresh graduate, open to full-time roles
- 📜 Google Data Analytics Professional Certificate (2026)
- 🏆 Finalist at IARC 2025 (ITS) for industrial automation
Data and ML
Software development
Computer vision
- YOLO Fungal Detection (BRIN): Built an instance segmentation pipeline to detect Curvularia leaf disease. Used YOLO11 and Roboflow for annotation, achieving 89.6% precision and a 0.807 mAP50.
- Cyclistic Bike-Share Analysis: Google Data Analytics capstone project. Cleaned and analyzed millions of trip records in BigQuery, then visualized the insights through interactive dashboards in Power BI and Tableau.
- LSPR Ethanol Sensor: Processed optical fiber sensor data by configuring Neural Networks (LSTM) to capture sequence patterns in stability tests. This optimization brought the MAPE down to 0.60% for halal product quality control.
- Physics Scientific Calculator: Desktop app built with C# (.NET 10, WinForms) for physics computations. It features a custom constant system where users can define and save their own physical constants.
| Repository | What I did | Status |
|---|---|---|
| DebugProbe/DebugProbe.AspNetCore | Fixed a bug where a negative MaxEntries value caused an infinite loop in the DebugEntryStore component. |
✅ Merged |
| DebugProbe/DebugProbe.AspNetCore | Improved ignore path matching accuracy to prevent false positives. | ✅ Merged |
| 3060s/FruityScale | Added dynamic color coding for scale match percentage badges based on the score value. | 🟡 In Review |
