Remote Speech ML Engineer - Real Ownership & Flexible Hours

SoundNativ; Colombia; A distancia; COP 90.000.000 - 150.000.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Competitive pay

Remote-friendly

Flexible hours

Ownership from week one

Tools provided

Descripción de la vacante

SoundNativ, a Colombia-based talent, is hiring to own phoneme recognition, pronunciation scoring and accent analysis pipelines from training to production. You will deploy models behind a Python/FastAPI scoring service for real-time results, and build robust data pipelines for audio and text.

We value a strong foundation in ML, production readiness and a user-centered mindset to avoid unfair scoring for speakers of diverse languages.

Formación

• 3+ years of training, evaluating and deploying ML models in production, ideally in speech or audio.
• Fluent in Python and PyTorch.
• Experience with ASR, forced alignment or phoneme recognition.
• Comfortable owning infrastructure: GPUs, inference latency and cost.

Responsabilidades

• Training and fine-tuning speech models (wav2vec 2.0, Whisper-style encoders and friends) for phoneme-level scoring and accent detection.
• Deploying them behind our Python and FastAPI scoring service for real-time and batch inference.
• Building evaluation sets that cover the first languages our learners speak, and tracking how scores hold up across them.
• Data pipelines for audio and text: cleaning, labelling, augmentation and versioning.
• Exploring where LLMs and speech-to-speech models can coach, not just score.

Conocimientos

Python

PyTorch

ML in production

Speech processing

Herramientas

wav2vec 2.0

Whisper-style encoders

GPUs

Inference latency

Descripción del empleo

SoundNativ, a Colombia-based talent, is hiring to own phoneme recognition, pronunciation scoring and accent analysis pipelines from training to production. You will deploy models behind a Python/FastAPI scoring service for real-time results, and build robust data pipelines for audio and text.

We value a strong foundation in ML, production readiness and a user-centered mindset to avoid unfair scoring for speakers of diverse languages.