Research Engineer - Inference

Sourced from aihiringboard.com

ElevenLabs is seeking a Research Engineer to join their research team, focusing on deploying and optimizing frontier AI models for production. The role involves owning the path from research checkpoints to serving infrastructure, ensuring high-performance, low-latency inference for real-time applications. You will build tooling and systems that enable researchers to ship new models safely and efficiently.

• Deploy state-of-the-art ML models to production environments
• Optimize inference performance including latency, throughput, and cost
• Build and tune high-performance serving systems for real-time streaming workloads
• Create tooling and infrastructure for rapid model deployment
• Profile and eliminate bottlenecks across the serving stack

• Experience deploying and serving ML models in production, preferably for latency-sensitive apps
• Strong engineering skills in GPU programming (CUDA, Triton) or inference optimization (TensorRT)
• Proficiency with serving frameworks such as vLLM or SGLang
• Ability to autonomously diagnose and solve complex engineering problems
• Portfolio of past projects, designs, or GitHub contributions demonstrating hard problem-solving

• Experience with quantization, distillation, KV-cache optimization, and batching strategies
• Background in custom kernel development
• Prior experience in audio or speech AI models