Research
Computational biology
My research in computational biology explores how machine learning can accelerate various tasks and solve complex problems in this field. I focus on developing AI/ML models, iterative pipelines and optimization frameworks for drug development, molecular design, and protein engineering.
Current research projects include:
- Evaluation of drug–target affinity prediction models using molecular graphs and protein language models (GitHub)
- Multi-objective latent space optimization of molecular and peptide properties under structural constraints
- Peptide sequence optimization with a mixture-of-experts genetic algorithm (GitHub)
- Protein function prediction and generative peptide optimization
Signal processing
I am currently pursuing a PhD in Electrical Engineering and Computer Science, where my research focuses on advancing acoustic signal processing through modern machine learning methods.
Here are some of the research topics I am working on:
- Developing advanced AI/ML techniques for spatial audio analysis and microphone array processing
- Sound source localization using message passing graph neural networks (GitHub) (Paper)
- Geometric deep learning for modeling spatial relationships and learning geometry-aware representations
Computer vision
I am also involved in research and development of computer vision methods for practical applications that span environmental monitoring, medical imaging, and industrial quality control.
Some of the projects I have work on:
- Development of a thermal imaging camera trap for detecting small animals using YOLO and OpenCV (GitHub)
- Feature extraction and engineering for machine learning surface crack detection on construction surfaces (GitHub)
- Automated 3D segmentation tool for renal calculus and organ segmentation on CT scans (GitHub) (Paper)
- Automated sanitary freshener quality control and defect detection using SAM and OpenCV (GitHub)
- Hemorrhagic lesion detection and segmentation on head CT scans using U-Net
- Non-invasive heart rate estimation using image processing and facial video analysis