Preference Model

We're hiring Low-Level Engineers to design and build RL environments that teach LLMs kernel development, hardware optimization, and systems programming. The goal is to create realistic feedback loops where models learn to write high-performance code across GPU and CPU architectures.

This is a remote contractor role with ≥4 hours overlap to PST and advanced English (C1/C2) required.

Requirements

Minimal Qualifications

• Strong Python (engineering-quality, not notebook-only)
• Production mindset (debugging, reliability, iteration speed)
• Clear understanding of LLMs, their current limitations
• Ability to meet throughput expectations and respond quickly to feedback

You may be a good fit if one of the following applies

• Deep understanding of memory hierarchies (registers, L1/L2/shared memory, HBM, system RAM) and their performance implications
• Threading models, synchronization primitives, and concurrent programming (warps, thread blocks, barriers, atomics)
• Cache coherence, memory access patterns, coalescing, and bank conflicts
• JIT compilation frameworks (e.g., Triton, JAX/XLA, TorchInductor, Numba)
• AOT compilation and optimization passes (LLVM, MLIR, TVM)
• Compiler and kernel frameworks such as CUTLASS, BitBLAS, or JAX/Pallas
• Modern C++, including templates, concurrency, and build systems
• Assembly-level programming and low-level optimization across GPU and CPU architectures (e.g., x86, ARM, NVIDIA Hopper, NVIDIA Blackwell)
• Debugging and optimizing GPU kernels using CUDA and/or HIP/ROCm
• Developing PyTorch custom operators, backend extensions, or dispatcher integrations (e.g., ATen, TorchScript, or custom backends)
• Customizing, extending, or optimizing c, including distributed inference workflows
• GPU communication libraries and collectives, such as NVIDIA NCCL, AMD RCCL, MPI, or UCX
• Mixed-precision and low-precision kernels (e.g., FP16, BF16, FP8, INT8), including numerical stability and performance trade-offs

We’re hiring RL Environments Engineers to design and build MLE/SWE environments that deliver high-quality, diverse tasks with minimal supervision. You will target a specific language model, meet a defined difficulty distribution, and deliver about one task every 10 hours. This is a remote contractor role with ≥4 hours overlap to PST and advanced English (C1/C2) required.

Responsibilities

• Design and build MLE/SWE environments and diverse tasks.
• Target a specified language model and satisfy the required difficulty distribution.
• Deliver ~1 task per 8-10 hours once onboarded.
• Edit tasks within 24 hours based on customer feedback.
• Onboard quickly and start delivering on day one with minimal supervision.
  
   

Requirements

What we’re looking for (must-haves)

• Strong Python (engineering-quality, not notebook-only).
• Hands-on LLM/GenAI work in production: you’ve shipped and operated real systems (not “wrapped an API and called it AI”).
• Strong product/engineering ownership: comfortable building, fixing, and scaling end-to-end pipelines.
• Docker + production mindset (debugging, reliability, iteration speed).
• ≥4 hours PST overlap and advanced English (C1/C2) for specs, reviews, and feedback.
• Ability to meet throughput expectations and respond quickly to feedback.
   

Strong signals (nice-to-have, big plus)

• Experience designing environments/tasks for RL and/or evaluations.
• Experience in high-stakes or regulated domains (e.g., healthcare, finance, fraud/risk, safety-critical systems).
• ML systems experience: CI/CD, monitoring, evaluation harnesses, MLOps, scalable pipelines.
• Systems depth: C++/Rust/Scala/Java, performance/infra optimization, distributed systems.
• Exposure to RL / bandits / agentic systems (not required, but a strong signal).
   

Not a fit if

• You’re primarily a prompt engineer without strong ML/engineering foundations.
• You’re a research-only / academic-only profile with little or no shipping/production ownership.
• You’ve only built in notebooks or rely heavily on managed AutoML tools.
  
   

Working Conditions

• Remote, independent contractor engagement.
   
• 40 hours/week - full time - need 4 hours overlap in the working hours with the team in Pacific time zone;
• Deliverables-driven; begin shipping on day one.

Conversion & relocation: Potential path to FTE and relocation to the Bay Area if performance and mutual fit align.