On the job market!
Seeking an ML for Science postdoc position, starting in Summer / Fall 2027.
Hello!
สวัสดีครับ (Sawasdee Krub) 🙏
I am a Ph.D. candidate in Applied Mathematics at the Center for Applied Mathematics, Cornell University, USA, where I work with Prof. Peter Frazier on Bayesian optimization and AI for Science.
I received a Bachelor’s degree in Mathematics (2017) and a Master’s degree in Applied Mathematics (2020) from Mahidol University, Thailand. My master’s thesis focused on Bayesian optimization for set-valued input functions under the supervision of Assoc. Prof. Tipaluck Krityakierne, with co-supervision from Prof. David Ginsbourger during my internship at Idiap Research Institute, Switzerland. I also obtained a second Master’s degree in Applied Mathematics from Cornell University in 2024.
I am a recipient of the DPST scholarship from the Institute for the Promotion of Teaching Science and Technology (IPST), Thailand, supporting my studies from undergraduate through doctoral levels.
My research focuses on Bayesian optimization, machine learning, and their applications to scientific problems, including food science, materials design, and chemistry.
Research Highlights
AI for Science
Materials Characterization
Developed scalable Bayesian optimization of composite functions for image-based inverse problems, estimating specimen thickness and mistilt from electron microscopy patterns to improve downstream ptychography reconstruction with far fewer expensive simulations.
AI for Science
Protein Function Prediction
Developed machine learning methods for protein function prediction under heavily biased evolutionary and surveillance data using positive-unlabeled learning and evolutionary modeling.
AI for Science
Formulation Design
Applied Bayesian optimization and machine learning to optimize protein formulation design for improved thermal stability and functionality in food systems.
Grey-box Bayesian Optimization
Function Networks & Partial Evaluations
Developed Bayesian optimization methods that exploit internal objective-function structure and partial evaluations, substantially improving sample efficiency for expensive scientific optimization problems.
Kernel Methods
Set-valued Input Functions
Developed kernel methods and Bayesian optimization techniques for optimization problems where the inputs are sets rather than fixed-dimensional vectors, enabling black-box optimization over combinatorial and structured domains.
Last updated: September 2, 2026