Senior AI Compiler Engineer - Applied Research
Core
Design and implement AI-based technology to solve core problems of low-level GPU programming and optimize compiler pipelines using reinforcement learning and prompt engineering.
Role type
Senior AI Research Engineer / Applied Scientist (Compilers)
Builds
AI-driven compiler solutions and optimization toolchains for NVIDIA's software stack
Domain
AI Systems / Compiler Engineering / GPU Optimization
Deliverable
production ML models
Required skills
Reinforcement Learning (policy gradients, actor-critic, offline RL), Large Model Training (Transformers, PEFT/LoRA, distributed training), Low-level Systems Programming (C++), Prompt Engineering, Python, CUDA Programming, Performance Profiling, Formal Methods/Static Analysis
Preferred skills
NVIDIA NeMo framework experience, GPU performance benchmarking, Distributed training/inference at scale
Technologies
Python, C++, CUDA, Transformers, PEFT, LoRA, RLHF
Responsibilities
Design AI-based solutions for low-level GPU programming, Build training pipelines for supervised fine-tuning and RL, Define model inputs/outputs over compiler representations, Develop evaluation frameworks for code quality and runtime, Prototype model architectures for scheduling and allocation, Create datasets from compiler traces, Apply RL to optimize performance and instruction-level parallelism, Integrate learned policies into production toolchains
Seniority
Senior, hands-on IC
