Personal research project index

Chenchen CHU（储辰辰）

I work at the intersection of AI-assisted research tooling, computational mechanics, additive manufacturing, and robust topology optimization. This page is a working map of research code, agent workflows, computational design experiments, and references I want to reuse.

GitHub · Google Scholar

Focus

• AI-assisted research tooling — Agent skills, patent-search tools, LaTeX templates, and structured workflows for research automation. agents MCP research workflows
• Computational design methods — Differentiable programming, FEM, homogenization, signed-distance fields, and topology optimization. JAX FEM topology optimization

Project index

A practical inventory of repositories I maintain or return to as references. Ownership labels distinguish my work from forks and external reference code.

Research & Patent Intelligence

Tools and notes for structured, LLM-assisted patent retrieval.

• Hello_Patent — Case-driven exploration of patent-search workflows using Espacenet and EPO resources. own research note patent search LLM workflow
• MCP_EPO — FastMCP tools for EPO OPS patent search, bibliographic lookup, abstracts, patent families, legal status, and number conversion. own tooling MCP EPO OPS Python

Research Productivity & Agent Skills

Reusable writing, proposal, and application workflows.

• dfg-wbp-latex-template — An anonymized LaTeX template for DFG Walter Benjamin Programme proposals, preserving structure, formatting, work-package logic, and writing scaffolds. own template LaTeX grant proposal
• job-application-tailor-skill — Agent skills and LaTeX templates for producing coordinated, evidence-grounded resumes and cover letters from a job description. own workflow agent skill resume
• grant-proposal-development-skills — Agent skills for call decoding, problem framing, work-programme design, reviewer simulation, compliance checking, and academic polishing. own workflow agent skill grant writing

Computational Design & Topology Optimization

Research code for implicit geometry, optimization, multiscale material design, and FEM-ready geometry pipelines.

• SDF-VAE — VAE workflow for implicit 3D structure parameterization using signed distance fields, including dataset preparation, CNN training, and STL reconstruction. own research code SDF VAE 3D structure
• jax_wrap_adjoint_homo — JAX-based thermo-mechanical homogenization and microstructure optimization using autodiff for unusual effective material properties. own research code JAX homogenization autodiff
• Jax_ProgMat — JAX and MMA workflow for programmable material-distribution design targeting prescribed deformation behavior. own research code JAX MMA
• sdf_shape_blending — Signed-distance-field shape blending for continuously varying microstructures generated from binary seeds. own geometry tool SDF shape blending
• voxel_composites_lattice_design — Voxel and lattice composite design scripts for tunable thermal expansion and Poisson’s ratio, with outputs for downstream homogenization. own design scripts voxel lattice design
• im2mesh4fem — Image-to-mesh workflow for FEM analysis, paired with FEniCS notes and simulation setup references. own utility notebook mesh generation FEM

FEM, Homogenization & Simulation References

Forked repositories retained as a practical numerical-methods toolbox.

• scikit-fem-jax — Fork of the lightweight scikit-fem finite-element assembly library for FEM prototyping and numerical experiments. fork / reference FEM Python
• FE2-Computations-With-Deep-Neural-Networks — Reference implementation for DNN-accelerated FE2 computations, including data generation and Fortran integration. fork / reference FE2 DNN
• fenics-wrinkle — FEniCS/MOSEK reference for reproducing wrinkle-related simulations and optimization plots. fork / reference FEniCS MOSEK
• HierarchicalDesign — Reference for architected-material design using VQ-VAE and diffusion models for forward and inverse microstructure problems. fork / reference VQ-VAE diffusion metamaterials
• TOuNN — Reference notebook for neural topology optimization, using network activations to parameterize the density field. fork / reference topology optimization neural networks

Learning & Reference Repositories

Learning material and notebooks retained for quick lookup.

• machine-learning-book — Forked notebooks for Machine Learning with PyTorch and Scikit-Learn, useful as a practical ML and teaching reference. fork / reference PyTorch scikit-learn notebooks