Haolin Wang
I'm Haolin Wang, a PhD candidate in Computer Science at the University of Sheffield (part-time).
About
I’m Haolin Wang, a PhD candidate in Computer Science at the University of Sheffield (part-time). I focus on developing machine learning methods to predict materials properties that are realistic and reliable, aiming to bridge the gap between model development and real-world materials discovery.
My work explores how multimodal and multi-fidelity data can be integrated to support reliable predictions, with emphasis on interpretability, deployment-centered evaluation, and uncertainty quantification. I strive to build automated pipelines that connect data curation, model training, benchmarking, reliability evaluation, and interpretation to enable realistic development, validation, and deployment of ML in materials science.
- 4 years' experience
- Joined April 2026
Skills· 13
Languages
- EnglishFluent
- ChineseFluent
Industry experience
- Software & Internet
- Education
- Professional Services
More experts in Sheffield, United Kingdom
Work experience· 6 roles
Implemented and benchmarked a diverse range of machine learning models, built evaluation pipelines and reliability metrics, and led the design of RealMat-BaG, a multi-fidelity materials dataset and benchmarking platform. Curated data sources to assess GNNs and ML baselines under domain-aware OOD scenarios; established model reliability through multi-level interpretation (gradient saliency, SHAP) to validate representations against physical principles. Maintained PyKale, engineered YAML-configurable training pipelines, and deployed public leaderboards to foster reproducible benchmarking. Supported planning and execution of the annual Multimodal AI workshop.
Helped identify transportation network risks, improved operational effectiveness, and provided business insights. Built automated dashboards (Tableau, Power BI) from diverse data sources using Amazon S3 and Redshift; supported project planning and delivery; automated a recurring email workflow in Python, saving 50+ hours per year.
Cleansed and analyzed industry data to identify market trends and opportunities; designed an automated operational dashboard using Tableau to provide weekly updates to the client’s leadership.
Education
Hire Haolin Wang today
Post a job and invite Haolin, or contact them directly. No hidden commission.