Research Engineer, Pre-Training
Core
Building massive-scale foundation models for financial market prediction and trading by developing fault-tolerant training infrastructure, data pipelines, and custom kernels.
Role type
Senior IC research engineer (pre-training)
Builds
Foundation models for trading signals
Domain
Financial markets + Large-scale ML systems
Deliverable
production ML models
Required skills
Large-scale distributed training optimization, custom kernel development, deep learning frameworks (PyTorch/JAX), high-performance computing, data pipeline engineering, mixed-precision training, model parallelism
Preferred skills
CUDA kernel development, Triton/Pallas/CuTe DSLs, XLA optimization, reinforcement learning, fine-tuning techniques
Technologies
GPUs, TPUs, NVLink, InfiniBand, Kubernetes, Slurm, PyTorch, JAX
Responsibilities
Build fault-tolerant infrastructure scaling across thousands of GPUs/TPUs, engineer data pipelines streaming terabytes per second, design custom kernels for efficiency gains, co-design novel architectures with researchers
Seniority
Senior, hands-on IC