Dynamic and results-driven researcher with expertise in machine learning and deep learning, honed at Michigan State University. Achieved a 90% success rate in targeted voice conversion while mentoring peers. Proficient in deep learning and passionate about advancing AI technologies through innovative solutions and collaborative teamwork.
Michigan State University, East Lansing, MI, Research Assistant, Developed targeted adversarial attack algorithms for voice recognition systems using Whisper and transformer models, creating imperceptible audio modifications that bypass human detection while remaining interpretable by voice assistants., Engineered an efficient adversarial audio generation pipeline with TensorFlow and PyTorch on a multi-GPU setup, enabling robust experimentation to identify vulnerabilities in automated voice systems and explore defense strategies., Michigan State University, East Lansing, MI, Research Assistant, Compared simulated gene trees with real ones using “Seq-Gen” and “Fast Tree” to develop Robinson Foulds distance for determining the similarity between two gene trees based on the Jukes Cantor Model, Poster Presentation: D. Dar and K. Liu, “Novel algorithms and tool development for comparative genomics and phylogenomics,” at 2021Mid-Michigan Symposium for Undergraduate Research Experiences (Mid-SURE) of Michigan State University, 2021. Video, Poster, Abstract