Jimming He

Hi! I'm an engineer and junior researcher with interests in multimodal alignment and generative modeling. I plan to apply for PhD programs, and in the meantime, I'm actively seeking opportunities for visiting research positions, research engineer roles, or residencies to further deepen my experience.

I graduated with my Master's and Bachelor's degrees in Computer Science from Stanford University in 2024. I was briefly involved in the Stanford Vision & Learning Lab led by Prof. Fei-Fei Li. In the past, I've interned at Meta AI, NVIDIA, Facebook, AFRL, Cleanlab, and Coursera.

I'm also interested in startups. I recently co-founded a company to build AI tools that automate and simplify GPU development. If you're interested in collaborating, please reach out.

Email / LinkedIn / Google Scholar

News

03/25 : Received funding from LeapYear to work on an exciting new project, alongside my friends William, Shafin, and Nathan.
12/24 : Graduated with a MS in Computer Science from Stanford University!
12/24 : Completed an internship at NVIDIA, where I researched LLMs for in-car applications.
09/24 : My first paper was accepted to NeurIPS'24!
09/24 : Completed an internship at Meta AI, where I developed safety classifiers for Llama 3 models.
06/24 : Graduated with a BS in Computer Science from Stanford University!
04/24 : Completed an internship at Cleanlab, where I showcased data curation tools for more reliable RAG and LLM fine-tuning.
04/24 : I'm a University Founder Fellow at Unusual Ventures!

Publications

(*) = Equal contribution

HourVideo: 1-Hour Video-Language Understanding

Keshigeyan Chandrasegaran, Agrim Gupta, Lea M. Hadzic, Taran Kota, Jimming He, Cristobal Eyzaguirre, Zane Durante, Manling Li, Jiajun Wu, Li Fei-Fei
NeurIPS Datasets and Benchmarks, 2024.
arXiv / project / teaser / code / openreview / poster / dataset

Projects

	How to detect bad data in your instruction tuning dataset for better LLM fine-tuning Jimming He, Sanjana Garg, Jonas Mueller Automatically catch low-quality responses, incomplete/vague prompts, and other problematic text (toxic language, PII, informal writing, bad grammar/spelling) lurking in any instruction-response dataset. blog / tutorial / code
	Audioformer: A Transformer-Based Approach for Audio Denoising Jimming He, Alex Kwon, Suhas Chundi poster
	3D U-Net Architectures for Multimodal Brain Tumor Segmentation Jimming He, Alexander Kwon, Pranav Gurusankar poster
	Deep Learning for Multi-Instrumental Music Generation Constance Horng, Jimming He, Ethan Ko poster