New Delhi : Building an AI prototype is one challenge. Running it reliably, securely and cost-effectively at enterprise scale is another. While AI adoption is widespread, translating experimentation into measurable business value remains difficult. The challenge increasingly lies in the infrastructure required to deploy, monitor, optimise and govern AI systems in real-world environments.
According to McKinsey’s The State of AI in 2026: On the Road to ROI, nearly nine in ten organisations regularly use AI, but only 44% report scaling it across the enterprise. Just 37% attribute any earnings before interest and taxes (EBIT) impact to AI, while one in five report that AI operating costs, including token expenses, constrain its use. Among large organisations, 40% report scaling AI agents, up from 27% a year earlier.
Addressing this gap, iHUB DivyaSampark, IIT Roorkee, in collaboration with Emeritus, has launched the Post Graduate Certificate in Agentic Systems & Production AI. Starting December 22, 2026, the 24-week, 100% online programme focuses on the engineering disciplines required to deploy, operate and maintain production-grade AI infrastructure.
From DevOps and MLOps to Production AI
DevOps transformed software delivery through automation and continuous integration. MLOps extended these practices to machine learning through experiment tracking, model versioning and lifecycle management. Production AI builds on both, addressing the infrastructure demands of large language models (LLMs), retrieval-augmented generation (RAG) systems and AI agents, including inference latency, computing costs, scalability, security and reliability.
Production AI engineering focuses on deploying, serving, monitoring, securing and optimising the underlying models, LLMs, RAG systems and agent runtimes at scale.
The curriculum brings together four interconnected disciplines:
MLOps: Automated ML pipelines, experiment tracking, model registries, CI/CD for ML, deployment controls and drift monitoring.
LLMOps: LLM hosting, inference serving, quantisation, prompt caching, token-cost modelling and managed-versus-self-hosted deployment trade-offs.
RAGOps: Vector databases, retrieval infrastructure and the monitoring and optimisation of RAG applications.
AgentOps: Agent runtimes, deployment, observability, security and operational controls for AI agent services.
These capabilities are supported by cloud-native engineering using Docker, Kubernetes and Terraform, alongside practices for high availability, disaster recovery, service-level agreements (SLAs), service-level objectives (SLOs), access control and audit logging.
Hands-on learning in production AI engineering
The programme combines live instruction with virtual labs and practical exercises, covering 25+ tools and platforms, including Kubernetes, Terraform, Airflow, MLflow, vLLM, KServe, LangGraph, Prometheus, Grafana and Milvus.
Six graded mini projects culminate in an Enterprise ModelOps and LLM Infrastructure capstone, integrating CI/CD pipelines, model registry management, drift detection, serving optimisation and cost governance into a production engineering project.
Participants develop the capabilities to deploy, scale, monitor, optimise and secure ML models, LLMs, RAG systems and AI agents. On successful completion, they receive a Post Graduate Certificate from iHUB DivyaSampark, IIT Roorkee.
Programme details
Start date: December 22, 2026
Duration: Six months (24 weeks)
Format: 100% online, with three-hour sessions every Saturday and Sunday
Learning commitment: 8–10 hours per week
Faculty engagement: Weekly sessions led by industry experts, supplemented by approximately 12 hours of live masterclasses by faculty members from IITs and other premier institutions
Eligibility: Bachelor’s degree, or a diploma with at least three years of work experience
Technical prerequisites: Intermediate Python and basic hands-on familiarity with cloud and DevOps tools, including AWS, Git, Docker, Kubernetes and CI/CD. This is not a beginner programme; applicants should already have a working foundation in Python and practical familiarity with cloud and DevOps environments.
Programme fee: INR 1,50,000 + GST
Optional campus immersion: Two days at IIT Roorkee’s Noida campus after successful programme completion
The programme is designed for DevOps, cloud, site reliability engineering (SRE) and platform engineers; software and backend engineers; ML, AI and data engineers; and technology architects, technical leads and engineering managers. Career Services, delivered through recorded content by Emeritus, is an additional programme feature.
“Organisations have moved quickly to experiment with AI, but running it reliably, securely and within budget is a different engineering challenge. The Post Graduate Certificate in Agentic Systems & Production AI is designed to help experienced engineers build the skills to deploy, monitor and govern AI systems in production. In collaboration with Emeritus, this 100% online programme combines hands-on learning, to strengthen the pool of engineers who can take AI from pilot to production,” said Manish Anand, Chief Executive Officer, iHUB DivyaSampark, IIT Roorkee.
****