Location: Remote (Americas) – Preference for São Paulo, Brazil or Medellín, Colombia

Compensation: $100,000 to $150,000 USD + Meaningful Equity

Job Type: Full-time, Remote

About the Role

We are looking for an Applied ML Engineer who thrives at the intersection of cutting-edge research and production-grade deployment. In this role, you will take full ownership of the machine learning models powering our clinical products, with a primary focus on developing next-generation Speech-to-Text (ASR) solutions for medical consultations in Brazilian Portuguese and Latin American Spanish.

This is a unique opportunity to solve a highly challenging domain problem where off-the-shelf ASR solutions fail. Leveraging our proprietary data advantage through our scribe product, you will build the data flywheels, infrastructure, and models that directly impact healthcare delivery.

Key Responsibilities

Own the Speech Stack: Manage the end-to-end speech-to-text pipeline, including data curation, training pipelines, evaluation frameworks, and production deployment.

Build the Data Flywheel: Transform scribe-generated transcripts into high-value training assets to continuously improve model accuracy.

Scale Infrastructure: Establish and maintain robust training infrastructure, including GPU orchestration and experiment tracking.

Expand Beyond ASR:



Collaborate with the AI team on LLM orchestration, fine-tuning small language models (SLMs), and building clinical classifiers.

Bridge Clinical & Tech: Partner with product and medical teams to translate real-world clinical requirements into robust model architectures and evaluation frameworks that reflect true clinical performance.

What We Are Looking For (Requirements)

Production-Proven Track Record: Deep experience operationalizing ML models in a production environment. This is an applied role, not a purely academic research position.

Speech & Audio Expertise: Hands-on experience building, training, and fine-tuning custom ASR, speech models, or audio ML from scratch (not just API integration).

Modern ML Stack Proficiency: Strong grasp of modern ML infrastructure (e.g., PyTorch, distributed training, data pipelines).

Autonomy & Grit: Ability to work independently and navigate ambiguous problems in a fast-paced environment.

Geographic Alignment: Must be located within the Americas time zones.

Job Type: Full-time

Application Question(s):

- ¿Tiene experiencia real en librerías de Audio/ASR (como Whisper/NeMo)?

Experience:

- Python: 3 years (Required)
- PyTorch: 1 year (Required)

Work Location: Remote

📌 Applied ML Engineer (Colombia)
🏢 Impulse Latam
📍 Colombia