Senior ML Engineer – Speech & Voice (NXJ-212)

The Role
GPU-based microservices handle the core speech pipeline—STT consumption, word-level alignment, diarization, and speaker identification. The primary challenge isn't just serving models, but establishing rigorous data-driven evaluation to separate real accuracy gains from benchmark noise on difficult sports audio. The Senior ML Engineer will own the speech services, lead fair model bake-offs, and hold full authority over which models reach production.
About the Product
The platform delivers real-time AI video processing and automated content generation for professional sports leagues globally. The underlying pipeline operates in high-noise live broadcast environments, demanding tight latency budgets, high throughput, and robust handling of overlapping speech and crowd noise.
Technology Stack: The platform delivers real-time AI video processing and automated content generation for professional sports leagues globally. The underlying pipeline operates in high-noise live broadcast environments, demanding tight latency budgets, high throughput, and robust handling of overlapping speech and crowd noise.
What You’ll Be Doing
-
Own and scale the core speech pipeline covering alignment, diarization, speaker identification, and enrollment signature-matching workflows.
-
Optimize GPU inference performance for throughput, memory footprint, and cost through batching, precision tuning, and compilation.
-
Build and maintain labeled sports test benchmarks stratified by speaker count, audio quality, language, and background noise.
-
Define and track production evaluation metrics including DER, WER, word-level speaker attribution, speaker-count error, latency, and GPU cost.
-
Conduct rigorous model bake-offs and write clear decision memos detailing trade-offs, licensing constraints, and statistical uncertainty.
-
Maintain a shadow or A/B deployment framework to gate every production release with empirical evidence.
What We Expect
Must-have
-
3–5+ years of experience shipping ML systems into production, with 2+ years dedicated to speech or audio pipelines.
-
Deep Python and PyTorch expertise, including hands-on GPU profiling, memory optimization, and inference acceleration.
-
Hands-on experience with at least two key speech domains: diarization (pyannote, NeMo/Sortformer, VBx), speaker embeddings (ECAPA-TDNN), or ASR and forced alignment (Whisper, Parakeet, wav2vec-family).
-
Strict evaluation rigor: demonstrated experience building test sets, computing DER/WER/EER correctly (handling collars and reference pitfalls), and running statistical hypothesis tests.
-
Demonstrated ability to critically analyze academic papers or model cards and reproduce published claims on internal datasets.
-
Hands-on experience deploying containerized GPU services in production environments using Docker and queue/API architectures.
Nice-to-have
-
Hands-on experience with Azure ecosystem tools (Service Bus, Blob Storage) and event-driven microservices.
-
Experience with domain adaptation or fine-tuning speech models on noisy, broadcast, or far-field sports audio.
-
Experience handling multilingual STT pipelines, particularly with Hebrew, Arabic, or Spanish.
-
Experience designing shadow deployment paths, experiment tracking workflows, and annotation team processes.
Why This Role Is Worth Your Time
-
Direct technical ownership of production GPU microservices where benchmark evaluations directly determine what reaches live users.
-
Applied research opportunity on challenging sports audio conditions (crowd noise, overlapping commentators, PA systems) rather than clean laboratory datasets.
-
Clear, evidence-based engineering culture where architectural and model choices are driven by data-backed decision memos rather than hype.