NVIDIA, GenAI Algorithms Intern – Model Optimization for Inference

In this role, you will design, implement, and productionize model optimization algorithms for inference and deployment on NVIDIA’s latest hardware platforms. The focus is on ease of use, compute and memory efficiency, and achieving the best accuracy–performance tradeoffs through software–hardware co-design.

Your work will span multiple layers of the AI software stack—ranging from algorithm design to integration—within NVIDIA’s ecosystem (Model Optimizer, NeMo/Megatron, TensorRT-LLM) and open-source frameworks (PyTorch, Hugging Face, vLLM, SGLang). You may also dive deeper into GPU-level optimization, including custom kernel development with CUDA and Triton.

What you’ll be doing:

• Help design and build modular, scalable model optimization software platforms that deliver exceptional user experiences while supporting diverse AI models and optimization techniques.
• Explore, develop, and integrate innovative deep learning optimization algorithms (quantization, speculative decoding, sparsity, distillation, pruning, neural architecture search) into NVIDIA's AI software stack.
• Deploy optimized models into leading open-source inference frameworks and contribute specialized APIs, model-level optimizations, and new features tailored to the latest NVIDIA hardware capabilities.
• Collaborate with NVIDIA teams on model optimization solutions for customer use cases, aiming for optimal end-to-end workflows and balanced accuracy–performance trade-offs.
• Conduct GPU kernel-level profiling to identify and capitalize on hardware and software optimization opportunities (e.g., efficient attention kernels, KV cache optimization, parallelism strategies).

• Pursuing a Master's or PhD in Computer Science, AI, Applied Math, or a related field.
• Strong deep learning background and hands-on experience with language models.
• Solid foundation in algorithms and programming.
• Proficiency in Python, PyTorch, and modern ML frameworks/tools.
• Good written and verbal communication skills, with the ability to work both independently and collaboratively in a fast-paced environment.

Ways to stand out from the crowd:

• Experience with model optimization and efficiency (e.g., quantization, distillation, pruning).
• Contributions to PyTorch, JAX, vLLM, SGLang, or other machine learning training and inference frameworks.
• Hands-on experience training or fine-tuning generative AI models on large-scale GPU clusters.
• Knowledge of GPU architectures and compilation stacks, and the ability to analyze and debug end-to-end performance.
• Familiarity with NVIDIA’s deep learning SDKs (Model Optimizer, NeMo).