AI/ML MLOps Engineer

Job Title: AI/ML
MLOps Engineer – LLM Fine-Tuning & Deployment

Experience: 5–8 Years

Key Responsibilities

• Fine-tune Large
  Language Models using Supervised Fine-Tuning (SFT) and Direct
  Preference Optimization (DPO).
  
• Develop and maintain training data pipelines, including data transformation, formatting,
  deduplication, filtering, and quality validation.
  
• Work extensively
  with the Hugging Face ecosystem, including Transformers, Datasets,
  and PEFT.
  
• Build and automate model
  evaluation and benchmarking frameworks to assess model quality and
  performance.
  
• Deploy and serve LLM
  models using AWS GPU/EC2 infrastructure and Amazon SageMaker.
  
• Optimize models for
  production through model quantization, inference optimization, and
  resource utilization.
  
• Build robust MLOps
  and ML CI/CD pipelines covering model training, evaluation, packaging,
  deployment, and monitoring.
  
• Implement A/B
  testing, Canary, and Shadow-mode deployments for safely introducing
  new model versions into production.
  
• Develop mechanisms
  for automated model promotion and rollback based on predefined
  performance and operational metrics.
  
• Implement production
  monitoring for model performance, latency, throughput, errors, GPU
  utilization, and resource consumption.
  
• Containerize ML
  workloads using Docker and deploy/manage them using Kubernetes/Amazon
  EKS.
  
• Collaborate with
  Data Scientists, ML Engineers, DevOps teams, and other stakeholders to
  build scalable and reliable AI/ML solutions.
  

Requirements

• Strong programming
  experience in Python.
  
• Hands-on experience
  with LLM fine-tuning, particularly SFT and DPO.
  
• Strong knowledge of Hugging
  Face Transformers, Datasets, and PEFT.
  
• Experience working
  with AWS GPU/EC2 and SageMaker for ML workloads.
  
• Strong understanding
  of MLOps, ML CI/CD, and model lifecycle management.
  
• Experience with LLM
  model serving and production deployment.
  
• Experience building training
  data preparation and processing pipelines.
  
• Knowledge of model
  evaluation, benchmarking, and performance optimization.
  
• Hands-on experience
  with model quantization.
  
• ​Experience implementing A/B, Canary, and
  Shadow-mode deployments
  

Benefits