Description

You will architect, manage, and scale the operational infrastructure for large language model workflows.

Responsibilities

• Design, build, and maintain LLM infrastructure including model serving, versioning, rollout, and inference pipelines.
• Automate model training, fine-tuning, evaluation, and retraining workflows.
• Implement observability, monitoring, logging, and alerting for model performance, drift, latency, and reliability.
• Optimize cost, scaling, caching, batching, and hardware utilization for inference.
• Lead MLOps best practices covering reproducibility, infrastructure as code, and model lineage.

Required Skills

• 5+ years of backend, MLOps, or infrastructure experience, including 2+ years with large language models or deep learning production systems.
• Deep familiarity with PyTorch, TensorFlow, Transformers, and Hugging Face.
• Experience with model serving platforms such as Triton, TorchServe, KFServing, or TensorFlow Serving.
• Strong knowledge of distributed systems, Docker, Kubernetes, and cloud infrastructure (AWS, GCP, Azure).
• Hands-on experience with orchestration tools like Airflow, Kubeflow, or MLFlow.
• Proven track record building scalable, low-latency inference systems.
• Familiarity with cost optimization strategies for model inference (quantization, pruning, batching, caching).
• Master’s or PhD in Computer Science, Machine Learning, or related field.

Key Skills

Education

Any Graduate