Senior AI Engineer – LLM Systems
Posted Updated
Description
Evollabs is building the next generation of AI infrastructure and seeks engineers focused on LLM systems, architecture, and performance. You will work at the intersection of model internals and hardware to ensure state-of-the-art models run efficiently on our platform.
Responsibilities include profiling, benchmarking, and optimizing LLM inference across distributed systems, designing attention and quantization optimizations, and collaborating with hardware teams to co-design efficient pipelines.
Analyze, profile and optimize LLM inference performance across distributed, multi-chip systems. Understand transformer architectures including dense and MoE models. Evaluate and benchmark LLMs on custom hardware (e.g., LLaMA, Mistral, Qwen, DeepSeek). Design optimizations for attention mechanisms (Flash Attention, grouped-query, sliding window). Work on model-level optimizations: quantization (INT8/FP8), KV-cache management, batching, parallelism strategies. Collaborate with hardware and compiler teams to co-design inference pipelines. Build and maintain benchmarking frameworks for latency, throughput, scaling. Analyze trade-offs between model architecture choices and system-level performance. Stay up to date with latest LLM research and inference optimization. Strong understanding of transformer architectures and LLM internals. Hands-on experience with multiple modern LLMs (e.g., LLaMA, Mistral, Qwen, DeepSeek). Deep knowledge of MoE architectures and attention optimizations.Skills
Sign in to read and join the discussion.