Ekfrazo Technologies - NLP Engineer - Machine Learning Solutions

About the Role :

We are seeking an experienced NLP Engineer with strong expertise in Azure Machine Learning to design, develop, fine-tune, and deploy NLP solutions for logistics and supply chain workflows.

The ideal candidate will have hands-on expertise in Hugging Face, spaCy, and Azure services, with proven experience in deploying scalable machine learning models in production Responsibilities (KRA) :

• Develop, fine-tune, and optimize NLP models for text classification, entity extraction, and intent detection using Hugging Face and spaCy.
• Build and deploy machine learning solutions on Azure ML and Azure Kubernetes Service (AKS) ensuring scalability, high availability, and low latency.
• Design and implement models for SKU classification, anomaly detection, and automated routing in logistics workflows.
• Develop data ingestion, preprocessing, and transformation pipelines integrating with APIs, warehouse management systems, and image recognition outputs.
• Monitor, evaluate, and retrain models using feedback loops, drift detection, and performance tracking via Azure Application Insights.
• Collaborate with cross-functional teams to integrate AI / ML capabilities into logistics operations and improve decision-making.
• Maintain best practices for MLOps, CI / CD pipelines, and automated deployments on Azure.
• Conduct research and stay updated on emerging NLP, Azure ML, and cloud-based AI innovations to continuously improve Skills & Qualifications :
• Proven experience (4 to 8 years) in Natural Language Processing (NLP) and Machine Learning.
• Strong expertise in Hugging Face Transformers and spaCy for NLP tasks.
• Hands-on experience with Azure Machine Learning for training, deployment, and lifecycle management of ML models.
• Experience deploying and scaling ML workloads on Azure Kubernetes Service (AKS).
• Proficiency in Python and ML frameworks (PyTorch, TensorFlow, or equivalent).
• Experience with data engineering workflows, API integrations, and preprocessing unstructured text data

(ref : hirist.tech)