cv

AI Automation Engineer

RedAwning

A leading vacation rental distribution and technology platform serving independent property managers worldwide.

• Sole technical lead for the company's AI initiative, owning end-to-end delivery: hardware selection, local inference server (vLLM/llama.cpp), backend integration, frontend design, and aligning direction with stakeholders up to CEO level.
• Architected LLM-based listing generation and a RAG-powered AI assistant, reducing per-listing creation time by ~90% across 100–250 properties daily.
• Built CI/CD infrastructure from scratch using Jenkins and Docker, reducing WebApp deployment to ~2 minutes and Airflow to ~30 seconds; cut incident recovery from days to under 5 minutes.
• Modernized engineering workflows and operational tooling across internal tools, Airflow pipelines, and EC2 batch jobs, introducing CI/CD standards, pipeline conventions, and documentation practices.

Algorithm Engineer

ByteDance

A global technology company behind TikTok, Douyin, and TikTok Shop.

• Integrated hundreds of thousands of product attributes using deep learning–based attribute matching and semantic search, reducing manual verification efforts by 15% and accelerating onboarding of new product categories.
• Standardized and optimized multimodal & multilingual attribute extraction workflows (LLM + Vision Models) across 5 languages, improving consistency of product listing quality and internal review efficiency.

Senior Data Scientist

Tokopedia

One of Indonesia's largest technology & e-commerce companies.

• Led end-to-end development of automated product categorization systems using deep learning and active learning, achieving >95% accuracy, 70% less labeled data, and contributing to a 1.3× uplift in transaction revenue from improved search & browse experience.
• Built and deployed BERT-based product attribute extraction across 5 parent categories with >90% accuracy, enabling processing of millions of product listings at scale.
• Designed a production-ready multimodal sensitive content/product classifier (Text + Image) with >90% accuracy, resulting in zero policy-violation escalation tickets during Ramadan, a critical peak period.
• Led a team of 5 to develop internal Active Learning and LLM-assisted labeling platforms, adopted by multiple teams and driving a 2× increase in model development productivity.

Software Engineer Intern

GDP Labs

A technology company specializing in AI, blockchain, cloud, and mobile/web platforms.

• Implemented a real-time on-the-fly subdomain provisioning feature for the company's microblogging platform, improving deployment flexibility for client demo environments.

Bachelor of Science

Universitas Gadjah Mada

Computer Science

Automated Active Learning for Large Scale Ecommerce Product Categorization

IP.com Technical Disclosure

Discusses the implementation of a novel automated active learning loop that can monitor model performance, use active learning for data collection, integrate with data labelling tools and trigger model re-trainings.

Multi-Modal Method To Identify Misrepresented Products In Ecommerce

IP.com Technical Disclosure

A multimodal approach to identify misrepresented products in e-commerce using text and image analysis.

Data Strategy For Large Scale E-commerce Product Categorization

IP.com Technical Disclosure

Presents data strategies for large-scale e-commerce product categorization systems.

1st Place, National Competitive Programming

Techphoria

1st place in the National Competitive Programming competition at Techphoria 2017.