Overview: Generative AI Course in Drumcondra

Built for Drumcondra builders

Every module ends with something working — a RAG pipeline, an agent, an eval suite — built live alongside peers from Drumcondra and beyond. This goes past prompting into the actual engineering behind production GenAI.

Curriculum breakdown

Module 1: Prompt engineering and structured chains

Drumcondra students learn structured prompt design and function calling, not just one-off prompt tricks.

Topics covered

• Prompt architecture and function calling
• Structured output validation
• Multi-step prompt chains

Module 2: RAG pipelines with vector databases

Building a working retrieval-augmented generation pipeline, end to end, as a Drumcondra student.

Topics covered

• Chunking and ingestion strategy
• Vector database selection and indexing
• Retrieval quality evaluation

Module 3: Agents and tool use

Building an agent that uses tools reliably, for Drumcondra students moving beyond single-turn prompting.

Module 4: Evals, guardrails and deployment

Shipping what you've built with monitoring and safety guardrails in place.

How the cohort runs

Format and schedule

Live, project-based cohort

Fully online, live over Zoom, for Drumcondra learners and peers across every module.

Who this course is for

Engineers and builders

Drumcondra engineers who want the architecture behind production GenAI, not just API basics.

Technical founders

Drumcondra founders building an LLM-powered product who need to move past a fragile prototype.

Outcomes and support

Portfolio project

Drumcondra students leave with a working RAG or agent project they can show employers.

Prerequisites

Basic Python recommended

The course focuses on architecture decisions for Drumcondra students, not syntax fundamentals.

Note

Covers OpenAI, Anthropic (Claude), and open-source models, so Drumcondra students aren't locked into one provider's mental model.

Quick answer

The Generative AI Course in Drumcondra is a hands-on, project-based program covering prompt engineering, RAG pipelines, agents and evals — the engineering behind production LLM systems, not just prompting tricks. It's built for Drumcondra engineers and technical founders who want to move a fragile prototype into something production-ready, and it's worth it if your goal is a working portfolio project, not just theoretical understanding of how LLMs work.

How this compares to free YouTube tutorials or API documentation

• You build a working RAG pipeline, agent and eval suite during the cohort, not just watch someone else build one
• Covers OpenAI, Anthropic (Claude) and open-source models together, so Drumcondra students aren't locked into one provider's documentation and mental model
• Live feedback on your architecture decisions from a practitioner shipping production GenAI systems, not a comments section
• Structured progression from prompting through RAG, agents, evals and deployment, instead of scattered standalone tutorials

Why Drumcondra builders choose Deepak Suhag

Taught by someone who ships production GenAI systems for clients — not a course built from documentation alone.