Machine Learning Intern - Multimodal Models Generative AI

NVIDIA NVIDIA · Semiconductors · STP, Hong Kong

NVIDIA is seeking a Machine Learning Intern to support research and development of large language and multimodal models. The intern will work on model fine-tuning, parameter-efficient training, architecture exploration, experiments, benchmarking, evaluation, data analysis, and prototype development using NVIDIA AI platforms and GPU-accelerated tools. The role also involves collaborating with researchers and engineers on AI innovation projects and exploring opportunities for technical publications.

What you'd actually do

  1. Support research and development of large language models and multimodal models.
  2. Work on model fine-tuning, parameter-efficient training, and architecture exploration.
  3. Assist with experiments, benchmarking, evaluation, and data analysis.
  4. Develop prototypes using NVIDIA AI platforms and GPU-accelerated tools.
  5. Collaborate with researchers and engineers on cutting-edge AI innovation projects.

Skills

Required

  • Pursuing BS, MS, or PhD in Computer Science, AI, Data Science, Engineering, Mathematics, or related fields.
  • Experience with machine learning / deep learning.
  • Strong Python programming skills.
  • Familiarity with PyTorch or TensorFlow.
  • Good analytical and problem-solving skills.
  • Good verbal and written communication skills in English.

Nice to have

  • Experience with LLMs, VLMs, multimodal AI, NLP, or generative AI.
  • Experience with distributed training or GPU computing.
  • Interest in applied research and publications.

What the JD emphasized

  • large language models
  • multimodal models
  • fine-tuning
  • parameter-efficient training
  • architecture exploration
  • experiments
  • benchmarking
  • evaluation
  • data analysis
  • prototypes
  • NVIDIA AI platforms
  • GPU-accelerated tools
  • technical publications
  • research outputs
  • LLMs
  • VLMs
  • multimodal AI
  • NLP
  • generative AI
  • distributed training
  • GPU computing
  • applied research

Other signals

  • multimodal models
  • large language models
  • fine-tuning
  • parameter-efficient training
  • architecture exploration
  • benchmarking
  • evaluation
  • data analysis
  • prototypes
  • NVIDIA AI platforms
  • GPU-accelerated tools
  • technical publications
  • research outputs