Harits Abdurrohman

Patented a cross-domain recommender using reinforcement learning and deployed regional Indonesian text-to-speech systems, improving TTS accuracy by 30% in three months and scaling solutions across Southeast Asia.

Developed medical imaging classification models, designed ETL pipelines for multi-source data, and implemented model versioning and monitoring to ensure reproducibility for healthcare clients.

Built video-to-annotation pipelines for autonomous vehicle perception datasets and co-authored an IEEE conference paper on dataset creation for AV perception.

Built data pipelines and dashboards for growth monitoring, developed churn prediction and user segmentation models, and delivered lifeplan back-testing simulations for financial products.

Researched computer vision for smart wheelchair navigation and published work on face sketch identification methods supporting academic research.

Activities and societies: Laboratory assistant research on computer vision for smart wheelchair navigation; published on face sketch identification methods.

Bachelor of Science in Computer Science focused on computer vision and image processing.

Completed AI Bootcamp program and was a Top 25 awardee.

Master of Science in Computer Science with focus on deep learning, NLP, knowledge distillation, and contrastive/self- and semi-supervised learning.