Senior Research Engineer

AssemblyAI
New York
Workplace: RemoteFull timeUSD 270,000 - 310,000 annuallyFunction: Research & Scientific (R&D)Skills: ["Measurement discipline","Communication","Collaboration","Skepticism","Cross-functional ownership"]

Build and evolve large-scale deep learning systems for voice AI, focusing on distributed training, data processing, and low-latency inference. Raise experimentation velocity by improving the end-to-end pipeline, maintaining JAX/TPU training frameworks, and creating evaluation and accuracy tooling. Optimize production inference (including quantization and speculative decoding), investigate performance bottlenecks across the stack, and translate research prototypes into production-ready systems through cross-functional collaboration.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
AssemblyAI
AssemblyAI
5 days ago

Senior Research Engineer

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 5 hours agoStatus: Live

Job Summary

Build and evolve large-scale deep learning systems for voice AI, focusing on distributed training, data processing, and low-latency inference. Raise experimentation velocity by improving the end-to-end pipeline, maintaining JAX/TPU training frameworks, and creating evaluation and accuracy tooling. Optimize production inference (including quantization and speculative decoding), investigate performance bottlenecks across the stack, and translate research prototypes into production-ready systems through cross-functional collaboration.
Location: New York
Workplace: Remote
Employment Type: Full time
Job Function: Research & Scientific (R&D)
Seniority: Mid level

Key Responsibilities

  • •Increase experimental velocity by making it faster to launch experiments, trust results, and decide what to try next.
  • •Maintain and scale the JAX training framework for large-scale distributed TPU training runs.
  • •Improve data quality for model training by investigating issues, building tooling, and turning findings into measurable accuracy gains.
  • •Build evaluation harnesses and analyze production model accuracy to identify improvements that matter most to customers.
  • •Translate research prototypes into production-ready systems and optimize production speech-model inference (architecture, quantization, speculative decoding), resolving bottlenecks across the stack.

Pay and Benefits

Salary: USD 270,000 - 310,000 annually

Key Requirements

  • •Expert-level proficiency with JAX and TPUs, including the surrounding ecosystem (Flax, Optax, and the XLA compilation pipeline).
  • •Strong experience optimizing inference systems for production, ideally with LLMs or speech models.
  • •Deep understanding of distributed training at scale and modern ML infrastructure best practices.
  • •Familiarity with inference optimization techniques such as continuous batching, KV-cache management, sharding, and quantization.
  • •Strong Python skills, with C++ or Rust experience a plus for kernel-level work.
Experience:Voice AIDeep learningDistributed trainingSpeech-to-textML infrastructureLLMs
Skills:Measurement disciplineCommunicationCollaborationSkepticismCross-functional ownership
Tech Stack:JAXTPUsFlaxOptaxXLAPallasPythonC++RustXLA compilation pipelineAsynchronous inferenceStreaming inferenceQuantizationSpeculative decodingContinuous batchingKV-cache managementShardingInference compilers

Company Brief

AssemblyAI
Builds developer-focused speech-to-text and audio intelligence APIs using deep learning. Offers transcription, speaker diarization, content moderation, sentiment and intent analysis, and other audio understanding tools for enterprises and developers to integrate voice capabilities.
Industry: API Platforms
Company Size: Medium (51 to 250 employees)
Growth: Growth Stage Startup
Headquarters: San Francisco, United States
Founded: 2017
WebsiteLinkedIn