Course Outline
Foundations of Agentic Systems in Production
- Agentic architectures: examining loops, tools, memory, and orchestration layers
- The agent lifecycle: covering development, deployment, and continuous operation
- Addressing the complexities of managing agents at production scale
Infrastructure and Deployment Models
- Implementing agent deployments in containerized and cloud-based environments
- Scaling strategies: comparing horizontal versus vertical scaling, concurrency management, and throttling
- Coordinating multi-agent orchestration and workload distribution
Monitoring and Observability
- Tracking key performance indicators: latency, success rates, memory utilization, and agent call depth
- Mapping agent activities and analyzing call graphs
- Implementing observability through Prometheus, OpenTelemetry, and Grafana
Logging, Auditing, and Compliance
- Establishing centralized logging and structured event aggregation
- Ensuring compliance and auditability within agentic workflows
- Constructing audit trails and replay mechanisms to facilitate debugging
Performance Tuning and Resource Optimization
- Minimizing inference overhead and streamlining agent orchestration cycles
- Utilizing model caching and lightweight embeddings to accelerate retrieval
- Conducting load testing and simulating stress scenarios for AI pipelines
Cost Control and Governance
- Analyzing cost drivers for agents: API consumption, memory, compute resources, and external integrations
- Monitoring agent-specific costs and applying chargeback models
- Enforcing automation policies to curb agent sprawl and eliminate idle resource consumption
CI/CD and Rollout Strategies for Agents
- Incorporating agent pipelines into CI/CD workflows
- Implementing testing, versioning, and rollback protocols for iterative agent updates
- Executing progressive rollouts and establishing safe deployment mechanisms
Failure Recovery and Reliability Engineering
- Engineering for fault tolerance and implementing graceful degradation
- Applying retry, timeout, and circuit breaker patterns to ensure agent reliability
- Developing incident response and post-mortem frameworks for AI operations
Capstone Project
- Developing and deploying an agentic AI system with comprehensive monitoring and cost tracking
- Simulating loads, assessing performance, and refining resource utilization
- Presenting the final architecture and monitoring dashboard to peers
Summary and Next Steps
Requirements
- A solid grasp of MLOps and production-grade machine learning systems
- Hands-on experience with containerized deployments using Docker and Kubernetes
- Proficiency in cloud cost optimization strategies and observability tooling
Intended Audience
- MLOps Engineers
- Site Reliability Engineers (SREs)
- Engineering managers responsible for AI infrastructure
Testimonials (3)
Strong confirmation that current LLMs are deployable for customer projects when the proper setup is in place.
Oliver Fromm
Course - Agentic AI for Enterprise Applications
Machine Translated
The trainer is patient and very helpful. He knows the topic well.
CLIFFORD TABARES - Universal Leaf Philippines, Inc.
Course - Agentic AI for Business Automation: Use Cases & Integration
The mix of theory and practice and of high level and low level perspectives
