Loading page...
Oktavian Andi Cahya Nugraha
I'm Oktavian, a frontend engineer interested in software engineering and AI. I build responsive web experiences, practical desktop applications, and machine learning projects that connect technical ideas with everyday needs.
Interests & focus
From ideas to working products.
Software Engineer
Web and desktop applications, API integration, and maintainable code.
Frontend Engineer
Responsive interfaces with React, Next.js, and TypeScript.
AI Engineer
Computer vision, natural language processing, and AI-powered features.
Central Java, Indonesia
About
I am Oktavian Andi Cahya Nugraha, a frontend engineer and master's student in Informatics Engineering at Dian Nuswantoro University. My interests span Software Engineering, Frontend Engineering, and AI Engineering, with a focus on usability, performance, and clean code.
My experience includes building SIMPP at PT PLN PUSMANPRO, developing personal web and desktop applications, and exploring computer vision and Indonesian sentiment analysis.
Current focus
Responsive interfaces, API integration, application state, practical software, and AI model design and evaluation.
Vision
Create software that is useful, maintainable, and easy to use, with AI applied to clear problems.
Experience
Nov 2024 - Oct 2026 | Semarang, Indonesia
PT PLN (Persero) Pusat Manajemen Proyek (PUSMANPRO)
Selected Work
Professional experience at PLN PUSMANPRO alongside personal projects in web development, desktop software, and applied AI.
Frontend engineering for PLN PUSMANPRO's SIMPP, featuring AI Reviews Evidence: field-level claim validation connected to supporting documents and highlighted source evidence.
An AI-powered CV review platform with ATS scoring, keyword analysis, and job description matching to help job seekers improve their resumes.
Website under maintenance
Visit Website (opens in a new tab)
A personal finance platform for tracking income, expenses, and cash flow, with interactive charts and receipt OCR to simplify transaction entry.
Website under maintenance
Visit Website (opens in a new tab)
An offline desktop point of sale application that helps small retailers handle checkout, manage inventory, and track sales without an internet connection.
An interactive portfolio presenting professional experience, personal software projects, and AI research through case studies and visual demonstrations.
A tourism platform helping visitors and local communities discover Kudus Regency, developed in collaboration with the Kudus Regency Culture and Tourism Office.
A website supporting health clinic operations, combining a Figma-designed interface with a native PHP implementation.
Fine-grained pavement crack segmentation using DAResUNet-ASPP: a Dual-Attention Residual U-Net with multi-scale context and boundary supervision.
A hybrid IndoBERTweet-BiLSTM model for sentiment classification in Indonesian tweets, reviews, and public opinions.
A comparison of convolutional autoencoders and U-Net for removing salt-and-pepper noise from PlantVillage tomato leaf images.
Skills
Applied AI
Computer Vision
DAResUNet-ASPP with residual blocks, dual attention, multi-scale context, and boundary supervision for fine pavement cracks.
Method: LOCO ablations, multi-seed validation, threshold sensitivity analysis
Datasets: Crack500, CrackForest, DeepCrack, non-crack images
Natural Language Processing
Combining IndoBERTweet contextual embeddings and BiLSTM sequence modeling for Indonesian sentiment classification.
Method: Hybrid modeling; accuracy, precision, recall, F1, and confusion matrix evaluation
Datasets: Indonesian tweets, reviews, and public opinions
Image Restoration
Comparing median filtering, convolutional autoencoders, and U-Net for restoring leaf images affected by salt-and-pepper noise.
Method: Median filter, CAE, CAE-SC, and U-Net comparison; PSNR and SSIM evaluation
Datasets: PlantVillage tomato leaf images
How I Work
Start with the task people need to complete, then make the interface clear and responsive.
Use reusable components, clear data flow, and thoughtful integration to make software easier to extend.
Connect model design with relevant metrics and experiments to understand what an approach can deliver.
Contact
Interested in Software Engineer, Frontend Engineer, and AI Engineer opportunities, as well as project and research collaborations.