General AI Specialist - PyTorch, MLflow, & Transformers at OneSeven Tech

Role overview Develop and productionize machine learning systems, with a focus on deep learning and Transformer models. The work combines model development and mathematical foundations with reproducible MLOps workflows and secure, performant model serving, moving solutions from research into practical delivery.

Responsibilities - Design and implement deep learning and Transformer models, comparing results with classical machine learning baselines. - Create reproducible experiment and model lifecycle workflows using MLflow and DVC. - Build, profile, and secure REST or gRPC APIs for production model serving. - Collaborate across disciplines in short, one-week delivery cycles to bring robust systems into use.

Requirements - Strong experience with PyTorch or TensorFlow/Keras and the Hugging Face ecosystem, including Transformers, Datasets, or Accelerate. - Advanced Python skills, including object-oriented and asynchronous programming. - Expert knowledge of MLflow Tracking, Projects, and Models. - Experience with advanced Git workflows and DVC. - Strong grounding in attention mechanisms, gradient descent, and linear algebra. - Proficiency with Docker and building secure model-serving APIs.

Nice to have - Familiarity with retrieval-augmented generation and vector databases. - Experience with MLflow Recipes, Kubernetes, distributed training, or model optimization tools such as DeepSpeed, ONNX Runtime, or TensorRT. - Hands-on experience with a managed cloud machine learning platform.

Benefits and work setup - Fully remote work from anywhere in Latin America. - Long-term contract beginning with a three-month period, with an expectation of moving to full-time afterward. - Paid time off and a referral bonus program.