Weijia He

• Led the development of a custom search engine leveraging customs data from 89 countries, integrating AI-enabled functions to support due diligence, client acquisition, market research, and strategic business decision-making.
• Recruited, trained, and managed an international trade operations team, including business development and order management personnel, to support overseas market expansion.
• Guided R&D strategy using data-backed market insights, translating customer demand, regulatory requirements, and competitive analysis into clear product development priorities.

Accelerating Materials Discovery Using ABACUS and DeePMD: A Computational Materials Science Project

• Dataset Preparation: Collecting a diverse dataset of known materials and generate training data by performing atomistic simulations with ABACUS, including atomic coordinates, energies, forces, and stress tensors.
• Machine Learning Model Development: Utilizing DeePMD to develop a deep learning model that accurately predicts atomic forces and energies. Training the model using the ABACUS-generated dataset, optimizing its performance.

• Led a team of programmers in overhauling the production line automation system, enabling a significant production upscaling from 500 tons to 3,000 tons of soy peptides per year.

• Collaboratively pioneered an innovative centrifugal spray dehydration method, establishing a new industry standard.
• Calibrated the ceramic microfiltration stage to enhance the precision of permeated peptides, resulting in a narrower weight range of 200-2000 Da, specifically targeting the desired size of 500-600 Da and improving overall yields.

• Improved soy peptide yield from 48% to 55% through meticulous experimentation and optimization of key proteolysis reaction conditions, such as temperature, pH level, and reaction time. Additionally, determined the optimal type, combination, and order of enzyme additions for maximum efficiency.