Member of Technical Staff - Research Engineer

Black Forest Labs
San Francisco, Freiburg
Workplace: HybridFull timeUSD 180,000 - 290,000 annuallyFunction: Research & Scientific (R&D)Skills: ["PyTorch","CUDA","Triton","CuTe","CUTLASS","NCCL","Nsight","Profiler","FP8","Quantization","Memory profiling"]

Embedded in production training to tackle performance, stability, and memory challenges for large multimodal models. You’ll optimize GPU kernels, enable low-precision training, profile training pipelines, and collaborate with researchers to translate architectural changes into efficient training implementations, owning the outcome and driving research-to-production progress.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Black Forest Labs
Black Forest Labs
2 months ago

Member of Technical Staff - Research Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 12 days agoStatus: Live

Job Summary

Embedded in production training to tackle performance, stability, and memory challenges for large multimodal models. You’ll optimize GPU kernels, enable low-precision training, profile training pipelines, and collaborate with researchers to translate architectural changes into efficient training implementations, owning the outcome and driving research-to-production progress.
Location: San Francisco, Freiburg
Workplace: Hybrid
Employment Type: Full time
Job Function: Research & Scientific (R&D)

Key Responsibilities

  • •Improve the performance, reliability, and numerical stability of production training runs for large multimodal generative models
  • •Profile full training steps across model code, attention, kernels, data loading, encoders, communication, optimizer steps, checkpointing, and memory pressure
  • •Implement and validate GPU-level optimizations: fused kernels, attention paths, low-precision matmuls, quantization kernels, CUDA/Triton/CuTe/CUTLASS experiments
  • •Push lower-precision training forward, including FP8 / MXFP8 / FP4-style paths, weight and activation quantization, accumulation choices, convergence risk
  • •Work with researchers to translate architecture changes into efficient training implementations, and help distinguish real model-quality progress from microbenchmark improvements

Pay and Benefits

Salary: USD 180,000 - 290,000 annually
Equity and Bonus:Equity
Perks:Equity

Key Requirements

  • •Strong PyTorch fluency with hands-on low-level training code experience
  • •Experience with distributed training concepts (FSDP, tensor/model/context/sequence parallelism, NCCL)
  • •Hands-on experience improving training throughput, memory footprint, or stability in real training runs
  • •Experience profiling GPU workloads with Nsight Systems/Compute or torch profiler
  • •Ability to work with researchers to translate architecture changes into efficient training implementations and diagnose training faults
Experience:Generative AIMultimodalDistributed training
Skills:PyTorchCUDATritonCuTeCUTLASSNCCLNsightProfilerFP8QuantizationMemory profiling
Languages:English
Tech Stack:PyTorchCUDATritonCuTeCUTLASSNCCLNsight SystemsNsight ComputeTorch profilerFP8MXFP8FP4Quantization

Company Brief

Black Forest Labs
Develops AI-driven solutions and provides machine learning consultancy to help businesses integrate advanced models and data-driven workflows. Focuses on prototyping, model deployment, and bespoke AI services to address industry-specific problems.
Industry: Consulting
Website