Intern- Cloud AI/ML Operations Engineer

Job responsibilities

• Assist in deploying ML and deep-learning models as inference endpoints on GPU instances and Kubernetes. Package models from Hugging Face into reproducible container images.
• Run and compare serving tools such as Triton Inference Server, vLLM, SGLang and Ollama. Record configuration, latency, throughput and GPU-memory usage so results can be reproduced.
• Support GPU instance setup, including NVIDIA drivers, CUDA, container runtimes and basic 
• Kubernetes GPU scheduling, following team procedures and senior engineering guidance.
• Monitor GPU utilization, memory, temperature and service health using existing dashboards. Collect logs, document anomalies and escalate issues to senior engineers.
• Perform guided Linux and networking tasks on cloud virtual machines, including checking users, services, ports, DNS and firewall configuration when troubleshooting connectivity.
• Support training and fine-tuning experiments by preparing datasets, running scripts and tracking runs with MLflow or a similar tool. Summarise benchmark and proof-of-concept findings.
• Help maintain Docker images and basic CI steps for ML workloads. Document setup procedures, operational checks and troubleshooting steps in runbooks and internal guides.
• Share progress and findings in team discussions, complete assigned learning activities, and apply feedback to improve the reliability and clarity of your work.

Skills

• Working knowledge of Python, with the ability to read, modify and debug scripts. Familiarity with Bash is useful for routine automation.
• Foundational understanding of ML or LLM workflows, including preparing inputs, running a model and interpreting evaluation results. Academic or project experience is sufficient.
• Hands-on exposure in any three of these areas is a strong starting point: Linux, networking, cloud infrastructure, Docker, classic ML, deep learning, LLMs and Hugging Face, model serving, or GPU computing.
• Ability to investigate a problem methodically, record commands and observations, and explain what was tried. Production experience or mastery of every listed tool is not expected.
• Clear written and spoken communication, careful documentation and a willingness to learn unfamiliar infrastructure with guidance from the team. 
• A GitHub profile or portfolio showing ML experiments, model deployments or automation. Experience with scikit-learn, PyTorch, TensorFlow or Hugging Face is useful.
• Experience building and running Docker images; familiarity with Kubernetes manifests or compute, storage and networking on AWS, Azure, GCP or AceCloud.
• Exposure to model-serving tools, MLflow, Prometheus or Grafana; awareness of batching, latency and throughput trade-offs, CUDA, GPU memory or accelerators.
• Relevant cloud fundamentals, Linux Essentials or NVIDIA Deep Learning Institute coursework or certifications. Certifications are optional.

Qualifications

• B. Tech degree in Computer Science (CS) or Information Technology (IT).
• 0-1 year experience.