[3 sessions] AI-Assisted Coding and Code Generation: Prompts to RAG, MCP and Agents

[3 sessions] AI-Assisted Coding and Code Generation: Prompts to RAG, MCP and Agents

We offer a 50% discount to students; please complete this form to avail the discount.

This three-session professional training programme in AI-assisted software engineering progresses from safe code generation on a single task to repository context through Retrieval-Augmented Generation (RAG) and the Model Context Protocol (MCP), and finally to agentic coding workflows.

The programme follows one continuous engineering discipline across all three sessions: bounded tasks, controlled context and verified output. Rather than treating each topic as a separate subject, participants apply the same core engineering principles throughout the course.

The sessions combine concise teaching with live demonstrations on a real sample repository, using Claude Code or Codex, Python, Git, pytest, Qdrant, the official MCP Python SDK, LangChain and LangGraph.

The Programme Consists of Three Sessions:

• Session 1 — Code Generation: Safe Prompts and Verified Change

• Session 2 — Repository Context: RAG and MCP

• Session 3 — Agentic AI: Agents, Workflows and Capstone

What You Will Learn

By the end of this programme, you will:

• Create bounded coding prompts with constraints, examples and acceptance criteria.

• Apply data classification and protect secrets and intellectual property.

• Verify AI-generated code using linting, pytest, security checks and human review.

• Decompose requirements into tasks with a clear definition of done.

• Debug failing tests, perform safe refactoring and review Git diffs.

• Build repository retrieval using embeddings, vector search and Qdrant.

• Improve retrieval quality through filtering, reranking, context budgets and evaluation.

• Build MCP servers exposing repository resources and typed tools.

• Apply least-privilege access, approvals, prompt-injection defence and audit logs.

• Design AI agents with typed tools, guardrails, tracing and stopping conditions.

• Model stateful coding workflows using LangGraph.

• Evaluate agent performance, security, trace evidence and cost.

• Build and verify a capstone code-generation assistant combining repository retrieval, MCP tools and a controlled workflow.

Why This Matters

AI-assisted software engineering requires more than generating code. Engineers must control the task, restrict the context and tools available to the AI, and verify the output before accepting a change.

This programme develops a repeatable engineering method that applies across individual coding tasks, repository retrieval, MCP tools and agentic workflows.

Ideal For

• Software engineers

• Technical leads

• QA engineers

• Software architects

• Final-year and postgraduate learners with programming and Git experience

Training Format

• 3 sessions

• 4 hours per session

• 12 hours in total

• Live online training

• 12 live demonstrations across the programme

• Four demonstrations per session

• A real sample repository used throughout the course

• A ten-minute break and five-minute quiz exactly at the two-hour mark of each session

• Course materials for all three sessions

• Quiz discussion in every session

• Sample repository, setup guide, lab instructions, governance checklists, testing checklists and handover checklists

Tools and Technologies Covered

Claude Code or Codex, Python, Git, pytest, Qdrant, the official MCP Python SDK, LangChain and LangGraph.

Requirements

Participants need basic programming ability, preferably in Python, and basic Git and command-line experience. A laptop capable of running the approved tools, access to Claude Code or Codex, and a stable internet connection are required.

Programme Outcome

Participants leave with a repeatable method for bounded and verified AI-assisted changes, progressing from a single coding prompt through repository retrieval and MCP tools to a verified agentic coding assistant with an operational handover pack.