Private Draft

The 29 personas behind AI

We’ve organized every stage and persona in the AI supply chain, informed by real recruiting at frontier companies. Click any row to see matching profiles from our talent graph.

Shaped by Industry Experts

← ATOMS & ENERGYUSERS & MARKETS →

← Back

Compilers & Frameworks

Toolchain under every training run

Compilers & Frameworks

Summary

Known as: Compiler Engineer, ML Framework Engineer, XLA/Triton Engineer, ML Compiler Engineer, MLIR Engineer, Kernel Engineer

Builds and maintains the frameworks, compilers, and runtime systems that researchers and engineers use to train and serve models. Owns the abstraction layer between model code and hardware: automatic differentiation, graph compilation, operator fusion, memory planning, and hardware-specific code generation.

Specializations

ML Frameworks — PyTorch, JAX, TensorFlow internals — graph representations, eager/compiled mode, autograd engines, distributed primitives. The programming model researchers actually write code against.

Compilers & Code Generation — XLA, Triton, MLIR, TVM — graph optimization passes, operator fusion, hardware-specific lowering, kernel generation. Turns high-level model descriptions into efficient hardware execution.

Runtime & Execution — Memory management, scheduling, multi-device orchestration, and the runtime that ties compilation to hardware execution. The layer that makes compiled graphs actually run efficiently on real clusters.

Where the Work Lives

[1]Substrate

[2]Compute

Primary

Graph compilation, hardware-specific code generation, and the runtime that ties compilation to execution.

[3]Intelligence

Secondary

Builds the frameworks (PyTorch, JAX) and autograd engines researchers write code against.

[4]Systems

[5]Distribution

Candidate Archetypes