AI/ML Engineer I — NLP, AWS, PyTorch

Akina Inc; Annapolis (MD); Sur place; USD 170 000 - 208 000; Plein temps

Avantages offerts par ce poste

PTO 24 days

Holidays 11

401k vesting 10%

Healthcare coverage

Education assistance

Résumé du poste

Akina Inc. in Annapolis Junction, Maryland seeks an Artificial Intelligence/Machine Learning Engineer I specializing in NLP, AWS, and PyTorch.

The role requires TS/SCI with polygraph within 7 years and involves designing, training, and deploying scalable AI/ML solutions across data pipelines and production systems. You will collaborate with data scientists, software engineers, and product teams to translate business needs into robust models, evaluate performance, and implement monitoring.

Qualifications

• Two years of applied ML experience required.
• Bachelor's degree in advanced math or CS required.
• Knowledge of NLP and language models preferred.
• Experience with AWS SageMaker, Git, and ML tooling.
• Ability to collaborate with data scientists and engineers.

Responsabilités

• Build, train, and fine-tune language models.
• Deploy models on AWS SageMaker.
• Design deployment pipelines with engineers.
• Evaluate model performance and iterate.
• Collaborate with data scientists and product teams.
• Document workflows and design decisions.

Connaissances

Language models

Marian NLP

Multilingual NLP

DL containers

GPUs

Model training

AWS SageMaker

PyTorch

TensorFlow

Kubernetes

C/C++

AI/ML frameworks

Formation

B.S. in advanced math/CS

Outils

AWS SageMaker

Docker

Git

MLFlow

Kubernetes

PyTorch

TensorFlow

Hugging Face

Description du poste

Akina Inc. in Annapolis Junction, Maryland seeks an Artificial Intelligence/Machine Learning Engineer I specializing in NLP, AWS, and PyTorch.

The role requires TS/SCI with polygraph within 7 years and involves designing, training, and deploying scalable AI/ML solutions across data pipelines and production systems. You will collaborate with data scientists, software engineers, and product teams to translate business needs into robust models, evaluate performance, and implement monitoring.