Senior Speech AI Researcher — Frontier Models & Impact

AssemblyAI, Inc.; United States; On-site; USD 120,000 - 230,000; Full time

Job summary

AssemblyAI, Inc. in the United States seeks a researcher to advance speech and multimodal language models. You’ll work with researchers and engineers to push frontier methods, build production-ready systems, and optimize inference stacks.

You will analyze model accuracy, surface data quality issues, and create tooling for performance analysis. The role emphasizes autonomy, deep technical focus, and the ability to own problems end-to-end.

Qualifications

• Hands-on experience training or adapting frontier speech or multimodal language models.
• Discipline in measurement and analysis.
• Skeptical of results until they hold up.
• Experience improving data quality and evaluating production models.
• Strong Python skills.

Responsibilities

• Research and develop new methods pushing the frontier of Speech AI.
• Improve the data our models learn from; surface quality issues and build tooling.
• Analyze the accuracy of production models and build evaluation harnesses.
• Partner with researchers and engineers to turn research prototypes into production-ready systems, modernizing model architectures and inference stacks.
• Find and fix performance bottlenecks across our training stack.

Skills

Python

Speech models

Model evaluation

Automation tooling

Autonomy

Job description

AssemblyAI, Inc. in the United States seeks a researcher to advance speech and multimodal language models. You’ll work with researchers and engineers to push frontier methods, build production-ready systems, and optimize inference stacks.

You will analyze model accuracy, surface data quality issues, and create tooling for performance analysis. The role emphasizes autonomy, deep technical focus, and the ability to own problems end-to-end.