Teams at KOG Co., better known as KOG Studios

Build AGCO, an agentic compiler for optimizing LLM execution across GPUs. Responsibilities include compiler and IR design, optimization passes, lowering, code generation, search-based implementation exploration, verification, GPU execution and profiling, LLM inference optimization, and measurement-driven optimization loops on real hardware. The role offers ownership across compilers, GPU systems, and inference, with the primary focus tailored to the candidate’s expertise.

Design and run experiments to measure how architecture decisions affect inference; create architecture variants optimized for inference speed; own post-training pipelines (fine-tuning, evaluation, adaptation); scale inference for large MoE models; publish research and build agent-driven research tooling.