Remote Senior Research Engineer, Large-Scale AI

Mosaec; United States; Remote; USD 270,000 - 310,000; Full time

Job summary

AssemblyAI is seeking a Senior Research Engineer to advance large-scale distributed training, data processing, and inference for state-of-the-art voice AI models. You will work within the Research team, collaborating across researchers, infrastructure, and production engineering to push model performance and efficiency.

You’ll optimize JAX and TPU workflows, implement improvements in model architectures, and contribute to production-ready systems.

Qualifications

• Expertise with JAX and TPUs, including Flax/Optax and XLA pipeline.
• Strong Python skills; C++ or Rust for kernel-level work is a plus.
• Experience optimizing inference systems for production, preferably with LLMs or speech models.
• Deep understanding of distributed training at scale and ML infrastructure best practices.
• Familiarity with modern inference optimization techniques: continuous batching, KV-cache, sharding, quantization.

Responsibilities

• Raise the team’s experimental velocity and help run reliable experiment jobs.
• Maintain and evolve the JAX training framework for large-scale TPU training.
• Improve data quality and build tooling to surface issues for measurable accuracy gains.
• Analyze production model accuracy, build evaluation harnesses, and prioritize impactful improvements.
• Translate research prototypes into production-ready systems and modernise architectures and infra.
• Optimize production inference for speech language models, including quantization and speculative decoding.
• Investigate bottlenecks across the stack from low-level kernels to high-level system design.
• Collaborate with researchers, infrastructure, and production engineering to trace problems to their source and ship fixes.

Skills

JAX

TPUs

Python

C++

Rust

Tools

Flax

Optax

XLA

Job description

AssemblyAI is seeking a Senior Research Engineer to advance large-scale distributed training, data processing, and inference for state-of-the-art voice AI models. You will work within the Research team, collaborating across researchers, infrastructure, and production engineering to push model performance and efficiency.

You’ll optimize JAX and TPU workflows, implement improvements in model architectures, and contribute to production-ready systems.