Developing a multimodal embedding-based retrieval over images, text, and graph information.

Designed an interview training system that combines facial emotion recognition.

Built an end-to-end pipeline for detecting and counting small biological objects.

Implemented a reproducible evaluation pipeline for DecodingTrust subsets.

Developed an attention-based deep learning model with 15% promote.

Started the master's program at Uppsala University.

Enhanced a ResUnet++-based watermark removal pipeline with attention and pooling modules.

Automated 3D content creation with Blender and Python and produced motion-driven visual assets.

Explored style transfer with frequency-domain decomposition and built controllable image-adjustment.

Integrated 3D image processing with LLM-assisted workflows to improve annotation efficiency and traceability.

Built a solid foundation in computer science through coursework and projects in programming, data structures, ML, and computer vision.