Amirali Aghazadeh

Georgia Tech

prof_pic.jpg

Coda S1209

Tech Square

Atlanta, GA 30308

I am an Assistant Professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology. I serve as program faculty for the Machine Learning, Bioinformatics, and Bioengineering Ph.D. programs. I am also affiliated with the Institute for Data Engineering and Science (IDEaS) and the Parker H. Petit Institute for Bioengineering and Bioscience. Prior to Georgia Tech, I was a postdoctoral researcher at Stanford and UC Berkeley, and earned my Ph.D. at Rice University.

My research lies at the intersection of machine learning and AI, information and signal processing, and biological design and engineering. I develop tools and algorithms that enable machines to learn, predict, scale, adapt, and be explained to solve most challenging biological science and engineering problems. Our solutions are usually enabled by a theoretical understanding of the underlying mathematical problems. For that reason, our group is a well-balanced mix of theory, algorithms, and real-world applications.

Some of the currect directions include:

  • Explainable, scalable, and safe AI4Science
  • ML/AI-guided protein design and optimization
  • Agentic AI for scientific discovery

I pronounce my first name /æmi:r’æli:/ and last name /ægə’zɑdɛ/.

news

Sep 29, 2025 Our LifeTracer work with NASA on discriminating Abiotic and Biotic organics in meteorite and terrestrial samples is now accepted to PNAS Nexus! 🎉
Sep 18, 2025 SHAP zero has been accepted to NeurIPS! 🎉 If you are looking for an ultra-scalable algorithm to explain your ML/AL sequence models and find long-range, high-order interactions, check SHAP zero.
Sep 18, 2025 SpecMER has been accepted to NeurIPS as a Spotlight! 🎉 We show that speculative decoding can be made even faster in autoregressive protein generation leveraging evolutionary information!
Jul 30, 2025 We won the IDEaS GenAI for Science competition! With support from Microsoft Azure, we will develop the next generation of agentic AI tools to accelerate scientific discovery.
May 02, 2025 Check out the story of our work on developing AI Scientists for hypothesis generation about Origins of Life in journal Nature. Many thanks to Celest Biever for covering this!

selected publications

  1. specme.png
    SpecMER: Fast Protein Generation with K-mer Guided Speculative Decoding
    Thomas Walton, Darin Tsui, Aryan Musharaf, and 1 more author
    Conference on Neural Information Processing Systems (NeurIPS) Spotlight (top 3% of 21,575 submissions), 2025
  2. SHAPzero.jpg
    SHAP zero Explains Biological Sequence Models with Near-zero Marginal Cost for Future Queries
    Darin Tsui, Aryan Musharaf, Yigit Efe Erginbas, and 2 more authors
    Conference on Neural Information Processing Systems (NeurIPS), 2025
  3. AstroAgents.png
    AstroAgents: A Multi-Agent AI for Hypothesis Generation from Mass Spectrometry Data
    Daniel Saeedi, Denise Buckner, José Aponte, and 1 more author
    International Conference on Learning Representation (ICLR) Workshop on Towards Agentic AI for Science: Hypothesis Generation, Comprehension, Quantification, and Validation, 2025