Agentic AI Engineering Bootcamp
The last course in the AI Engineer track. You move from single-shot RAG to agents that plan, act and observe: tool calling, the Model Context Protocol (MCP) for standard tool and data connections, multi-agent orchestration and agentic RAG. Then you take it to production with deployment, database scaling, security and observability, and ship a full agentic system.
- Duration
- 12 weeks · 60 hours
- Format
- Live online + self-study
- First class
- Free, no obligation
- Level
- Advanced (after AIE-2)
- Schedule
- Evenings, US time zones
- Lectures
- 21 (notes · lab · quiz)
- Certificate
- Advanced Professional Certificate
- Tuition
- $500 per student, batch of 3 to 5 · payment plans
What you'll master.
Foundations of Agentic AI
- Agents vs. chatbots vs. RAG
- The ReAct loop
- Agent memory and reasoning strategies
Tool Use at Scale
- Tool schemas and routing
- Errors, retries and fallbacks
- Sandboxing and permissions
Model Context Protocol (MCP)
- Hosts, clients and servers
- A custom MCP server for PostgreSQL
- MCP vs. plain function calling
Agent Frameworks & Orchestration
- Graph and state-machine orchestration
- Role-based frameworks
- Coordinator and worker patterns
Multi-Agent System Design
- Splitting work across agents
- Shared memory and state
- Human-in-the-loop and tracing
Agentic & Advanced RAG
- Retrieval as a tool
- Query rewriting and self-correction
- Hybrid and graph RAG
Databases & Infrastructure at Scale
- Pooling, replication and partitioning
- Multi-tenant schemas
- Query and semantic caching
Deploying Agentic Apps
- Docker
- Streaming, async and long-running APIs
- Task queues and workers
Security, Observability & Guardrails
- Prompt injection and tool misuse
- PII and data leakage
- Tracing and continuous evaluation
The complete 21-lecture path.
Every lecture includes a live class, self-study notes, a hands-on coding lab and a 10-question quiz.
01Agents vs. Chatbots vs. RAG
- What makes a system an agent (Module 1)
- Agents vs. chatbots vs. RAG pipelines (Module 1)
- The plan, act and observe (ReAct) loop (Module 1)
- Recap: tool calling and RAG from AIE-2
LAB Write a minimal ReAct loop with one tool
02Agent Memory & Reasoning
- Short-term memory: the context (Module 1)
- Long-term memory in a vector store or database (Module 1)
- ReAct, Reflexion and self-critique (Module 1)
- When an agent should stop
LAB Add database-backed memory to your agent
03Designing Tools & Routing
- Robust tool and function schemas (Module 2)
- Tool routing and selection (Module 2)
- Naming and describing tools for a model
- Testing tool choice
LAB Build a tool-calling agent that queries PostgreSQL
04Tool Errors, Retries & Safety
- Error handling and retries (Module 2)
- Fallback strategies (Module 2)
- Sandboxing tool execution (Module 2)
- Permissions on tools (Module 2)
LAB Make the agent recover from failing and slow tools
05MCP Architecture
- What MCP standardizes (Module 3)
- Hosts, clients and servers (Module 3)
- MCP vs. plain function calling (Module 3)
- Tools, resources and prompts in MCP
LAB Connect an agent to an existing MCP server
06Building MCP Servers
- A custom MCP server exposing PostgreSQL queries (Module 3)
- Exposing an external API (Module 3)
- Connecting to several MCP servers (Module 3)
- Testing an MCP server
LAB Build an MCP server with database and API tools
07Graph-Based Orchestration
- State machines and graphs, LangGraph style (Module 4)
- Nodes, edges and state
- Branching and loops
- Checkpointing a run
LAB Rebuild your agent as a graph with explicit state
08Role-Based & Hierarchical Agents
- Role-based frameworks, CrewAI and AutoGen style (Module 4)
- Coordinator and worker patterns (Module 4)
- Hierarchical agents (Module 4)
- Agent-to-agent messages and shared state (Module 4)
LAB Compare a coordinator design with a role-based crew
09Designing Multi-Agent Systems
- Splitting a complex task across agents (Module 5)
- Shared memory and state across agents (Module 5)
- Choosing how many agents you need
- Cost of extra agents
LAB Build a researcher, analyst and writer system with a coordinator
10Human-in-the-Loop & Tracing
- Human checkpoints and approvals (Module 5)
- Debugging multi-agent runs (Module 5)
- Tracing each step and tool call (Module 5)
- Replaying a failed run
LAB Add an approval step and trace a full run
11Retrieval as a Tool
- Retrieval as a callable tool (Module 6)
- Query rewriting (Module 6)
- Self-querying retrieval (Module 6)
- When the agent should skip retrieval
LAB Give your agent retrieval with query rewriting
12Self-Correcting & Graph RAG
- Iterative and self-correcting retrieval loops (Module 6)
- Hybrid and graph-based RAG (Module 6)
- Agentic RAG on pgvector at scale (Module 6)
- Measuring answer quality
LAB Build a self-correcting RAG loop over PostgreSQL and pgvector
13PostgreSQL at Scale
- Connection pooling (Module 7)
- Replication basics (Module 7)
- Partitioning (Module 7)
- Reading load metrics
LAB Load-test a database under concurrent agent traffic
14Multi-Tenant Data & Caching
- Multi-tenant schema design (Module 7)
- Vector index tuning for production (Module 7)
- Query and semantic caching (Module 7)
- Cost and latency trade-offs (Module 7)
LAB Add a semantic cache and measure the savings
15Containers & Agent APIs
- Containerizing an agent app with Docker (Module 8)
- Streaming and async APIs (Module 8)
- Long-running tasks (Module 8)
- Configuration and secrets in containers
LAB Containerize and deploy an agent with a streaming API
16Workers, Queues & Hosting
- Task queues and background workers (Module 8)
- Running multi-step agent jobs (Module 8)
- Managed vs. self-hosted model and agent infrastructure (Module 8)
- Rollbacks
LAB Move long agent runs onto a background worker
17Agent Security
- Prompt injection defenses (Module 9)
- Tool misuse defenses (Module 9)
- PII handling (Module 9)
- Preventing data leakage (Module 9)
LAB Add a prompt-injection defense to a deployed agent
18Observability & Continuous Evaluation
- Logging agent decisions and tool calls (Module 9)
- Tracing and monitoring (Module 9)
- Continuous evaluation in production (Module 9)
- Alerting on bad behavior
LAB Add logging, tracing and an evaluation job to your agent
19Capstone: Architecture & MCP Tools
- Architecture design (Capstone)
- Custom MCP tool build (Capstone)
- Data model in PostgreSQL and pgvector (Capstone)
- Instructor design review
LAB Design the capstone and build its MCP tools
20Capstone: Integration & Deployment
- Multi-agent integration (Capstone)
- Container deployment (Capstone)
- Monitoring and tracing setup (Capstone)
- Instructor check-in
LAB Deploy the capstone with monitoring
21Capstone: Security Review & Demo
- Security review (Capstone)
- Architecture and security report (Capstone)
- Final demo (Capstone)
- Next steps after the track
LAB Present the capstone and its written report
Three courses, one path.
Each course builds on the one before it, so you can take them back to back as one program.
- Course 1AI Engineering FoundationsBeginner · 4 weeks · 40 hours
- Course 2Generative AI EngineeringIntermediate (after AIE-1) · 8 weeks · 56 hours
- Course 3Agentic AI EngineeringAdvanced (after AIE-2) · 12 weeks · 60 hours
Who it is for
Developers who finished Generative AI Engineering (or have the same skills) and are aiming for senior AI engineer roles
Prerequisites
Generative AI Engineering (AIE-2), with solid Python and PostgreSQL.
Tools you will use
What you will ship
A production-grade multi-agent app (such as a research or customer-support agent) with custom MCP tools, a PostgreSQL and pgvector backend, container deployment, tracing and a written architecture and security report.
Assessment
Graded labs (35%), capstone demo, architecture write-up and security review (65%)
Certificate
Advanced Professional Certificate: Agentic AI Engineering, issued by Netwisers when you pass.
By the end you will
- Implement the plan, act and observe loop with short and long-term memory
- Build custom MCP servers that connect agents to PostgreSQL and external APIs
- Design multi-agent systems with clear roles and coordination
- Scale PostgreSQL for multi-tenant, high-traffic AI apps
- Deploy, monitor and secure an agentic app in a production-like environment
Roles this prepares you for
Includes resume review, LinkedIn optimization and mock interviews focused on the US job market.
Ready to start Agentic AI Engineering?
Book a free consultation. We will confirm the next start date in your time zone and walk you through tuition and payment plans.