Hi, my name is

Qi Huang.

I build AI for materials science.

I'm a postdoctoral researcher at the Shanghai Institute of Microsystem and Information Technology (SIMIT), Chinese Academy of Sciences, working at the intersection of deep learning, physics-informed modeling, and semiconductor materials.

View Full CV

About Me

Qi Huang

I am a postdoctoral researcher at SIMIT, CAS, working with Prof. Wenjie Yu under the Shanghai Super Postdoc Program. I received my Ph.D. in Microelectronics and Solid-State Electronics from SIMIT, CAS in 2025.

Prior to that, I earned an M.Sc. in Computer Science and Artificial Intelligence from the University of Manchester (2021) and a B.Sc. from the University of Nottingham Ningbo China (2020).

From July 2024 to March 2025, I was a visiting researcher at the RIKEN Center for Advanced Intelligence Project (AIP), Japan, collaborating with Dr. Qibin Zhao.

Here are some technologies I work with:

Deep Learning Physics-Informed ML Graph Neural Networks Molecular Dynamics Python PyTorch Knowledge Graphs LLM Agents Cheminformatics Multimodal Learning

Research

01 // Materials Representation

How can materials be understood by machines?

Materials carry intrinsic properties across modalities and scales that pure data-driven approaches struggle to capture. My work integrates physical mechanisms into representation learning — fusing molecular sequences, graph structures, 3D conformations, and domain knowledge into unified multimodal representations. Physical priors are embedded in downstream tasks such as neural network force fields, pursuing representations that are both interpretable and generalizable.

02 // Intelligent R&D Pipelines

How can materials R&D be intelligently driven?

Materials development involves a long decision chain spanning literature retrieval, data integration, multi-scale simulation, and process optimization. My work places intelligent agents at the center, orchestrating knowledge graphs, cross-scale simulation tools, predictive models, and process optimizers into an automated pipeline — with integrated circuit materials as the primary application domain.

Experience & Education

2025 — Now

Postdoctoral Researcher

Shanghai Institute of Microsystem and Information Technology, CAS · Shanghai, China

Working with Prof. Wenjie Yu on AI-driven materials research. Awarded the Shanghai Super Postdoc Program.

AI for Science IC Materials Intelligent Agents
Jul 2024 — Mar 2025

Visiting Researcher

RIKEN Center for Advanced Intelligence Project (AIP) · Japan

Collaborated with Dr. Qibin Zhao on multimodal representation learning for polymer property prediction.

Multimodal Learning Large Language Model
2021 — 2025

Ph.D. — Microelectronics & Solid-State Electronics

Shanghai Institute of Microsystem and Information Technology, CAS

Doctoral research on physics-informed ML for materials science. Developed neural network force fields, coarse-grained simulation methods, and multimodal polymer representations.

Deep Potential Material Representation Polymer ML
2020 — 2021

M.Sc. — Computer Science & Artificial Intelligence

University of Manchester · Manchester, UK
Machine Learning NLP
2016 — 2020

B.Sc. — Computer Science

University of Nottingham Ningbo China · Ningbo, China

Publications

Unified multimodal polymer representation framework
J1

Unified multimodal multidomain polymer representation for property prediction

npj Computational Materials, vol. 11, Art. no. 153, 2025

A multimodal prediction framework fusing SMILES, molecular graphs, descriptors, and 3D conformations with domain knowledge text, achieving state-of-the-art performance on multiple polymer benchmark tasks.

DOI: 10.1038/s41524-025-01652-z
Hierarchical deep potential coarse-grained modeling
J2

Hierarchical deep potential with structure constraints for efficient coarse-grained modeling

Journal of Chemical Information and Modeling, vol. 65, no. 7, pp. 3203–3214, 2025

A coarse-grained molecular dynamics method integrating Boltzmann inversion structural priors with hierarchical neural network potentials, achieving high structural fidelity while improving computational efficiency by an order of magnitude.

DOI: 10.1021/acs.jcim.4c02042
Weighted-chained-SMILES copolymer framework
J3

Enhancing copolymer property prediction through the weighted-chained-SMILES machine learning framework

ACS Applied Polymer Materials, vol. 6, no. 7, pp. 3666–3675, 2024

A machine learning framework based on weighted-chained-SMILES for copolymer property prediction, effectively encoding monomer sequence information and composition ratios to improve accuracy for copolymer systems.

DOI: 10.1021/acsapm.3c02715
Polyamine CMP silicon polishing simulation
J4

Mechanism exploration of the effect of polyamines on the polishing rate of silicon CMP

Nanomaterials, vol. 14, no. 1, Art. no. 127, 2024

A combined simulation and experimental study exploring how polyamine additives affect the silicon removal rate in CMP, revealing adsorption mechanisms and providing guidance for slurry formulation design.

DOI: 10.3390/nano14010127
P1

一种基于多源数据的聚合物统一表征向量生成方法

Chinese Invention Patent (pending) · Application No. 202511203392, 2025

First inventor. A unified polymer representation vector generation method based on multi-source data.

P2

抛光材料去除率分布预测模型训练方法、预测方法、存储介质和终端

Chinese Invention Patent (granted) · Patent No. ZL202410828607.2, 2024

Second inventor. Training and prediction method for polishing material removal rate distribution prediction models.

View All Publications

What's Next?

Get In Touch

I'm always open to discussing research collaborations, academic opportunities, or AI for Science projects. Feel free to reach out — I'll do my best to respond promptly.

Say Hello
GitHub ORCID Academic Page huangqi@mail.sim.ac.cn