At Monash University Indonesia, I developed a multitask deep learning model that uses RGB-D forest imagery to segment tree instances, classify species, and estimate trunk diameter. My thesis showed that depth improved segmentation AP50 by up to 16.93 percentage points across six backbone architectures, with a manuscript in preparation for submission to Ecological Informatics.
Before my data science work, I worked in internal audit, continuous improvement, environmental operations, and planning survey roles in the forestry industry. At Arara Abadi Indonesia (Djarum Forestry), I assessed internal controls and risk management systems; at Satria Perkasa Agung (Sinarmas Forestry), I worked on operational planning, GIS mapping, and forest sustainability and timber-legality certification records.
