Le Chen
AI Researcher at Argonne National Laboratory
Host, TPC Seminar Series
💡 LLM, HPC, AI4Code
✉️ lechen[AT]anl.gov
💻 DSL, Argonne National Laboratory
🌆 Chicagoland, Illinois
Le Chen is a Postdoctoral Researcher at Argonne National Laboratory, supervised by Dr. Rajeev Thakur. He received his Ph.D. in Computer Science from Iowa State University, advised by Dr. Ali Jannesari.
His research bridges high-performance computing and artificial intelligence along two complementary directions: HPC4AI, which develops scalable, efficient methods for training and inference of large AI models on modern supercomputing systems; and AI4HPC, which focuses on large language models (LLMs) for scientific code generation, translation, and reasoning.
Additional broad research interests include AI agents and AI4Science. His work has appeared in leading HPC and AI venues, including SC, NeurIPS, ACL, EMNLP, MLSys, IPDPS, NAACL, and EuroPar.
Speak at the TPC Seminar Series
Le hosts the TPC Seminar Series, a community venue under the Trillion Parameter Consortium that brings together researchers from academia, national labs, and industry working on large-scale AI — spanning foundation model training and inference, agentic systems, scientific applications, evaluation, and AI infrastructure at the trillion-parameter scale.
If you’d like to share your recent work with this community, please feel free to reach out — we welcome a broad range of topics across AI.
news
| Sep'26 Argonne | 🏆 Received an Impact Argonne Award in the category of Innovation for creating PETScAgent-Bench — grateful to the whole PETSc team!🎉 |
|---|---|
Sep'26  | 🎉 Our Genesis Mission Phase II project “AI4HPC: An Iterative Framework for AI-Assisted Scientific Software Development and Optimization” has been selected for award!✅ — honored to be part of it, and many thanks to the lead PI and the whole team! |
| Sep'26 NeurIPS'26 | 📄 One paper accepted at NeurIPS 2026 main conference!✅ See you in Atlanta 🌆! on why reviewer precision does not translate into critique uptake in multi-agent math reasoning |
| Sep'26 SC'26 | 📄 One paper accepted at SC workshop: AgenticAI4HPC!✅ on agentic CUDA-to-CSL kernel translation and optimization for wafer-scale systems |
| Sep'26 SC'26 | 📄 One paper accepted at SC workshop: AI4S!✅ on an agentic framework for evaluating AI-generated scientific code in PETSc |
| Aug'26 EMNLP'26 | 📄 One paper accepted (EMNLP Findings)!✅ on latent reasoning guidance for parallel code translation |
| Aug'26 ERAD'26 | 🌍🎤 Our work “LARS: LLM-assisted Radar Scenes” was presented at ERAD 2026 in Belgrade 🇷🇸! |
| Jul'26 | 🎉 Two of our Genesis Mission Phase I proposals accepted!✅ — grateful to the lead PIs and the whole team! Explore the Genesis Mission awards dashboard 📊 for a look at the full landscape. |