IICTI Certified Prompt Engineering Professional

Holders demonstrate that they can apply and evaluate lLM behaviour, tokenisation and decoding mechanics, advanced prompt architecture and instruction hierarchies, reasoning elicitation and task decomposition and structured outputs, tool use and function calling, plus 7 further areas, in realistic professional scenarios, as defined in the AQSCERT scheme for Prompt Engineering Professional.

Who it is for: Professionals working, or preparing to work, in artificial intelligence roles.

1. 01LLM behaviour, tokenisation and decoding mechanics5% of the exam

   • Evaluate how model architecture and decoding settings shape output quality and select parameters for a given task.
   • Scope: Transformer and attention basics as they affect prompting, tokenisation effects, context window limits and position effects, sampling parameters (temperature, top-p), reasoning models and thinking budgets.
2. 02Advanced prompt architecture and instruction hierarchies14% of the exam

   • Design robust prompt architectures with clear instruction hierarchies for production applications.
   • Scope: System vs user vs tool message roles, instruction precedence, structured prompts with XML/Markdown delimiters, role and context framing, prompt templates and versioned prompt libraries.
3. 03Reasoning elicitation and task decomposition14% of the exam

   • Design prompting strategies that improve reasoning reliability on complex multi-step tasks.
   • Scope: Chain-of-thought and extended thinking, few-shot and exemplar selection, self-consistency, prompt chaining, plan-and-execute, critique and self-refinement loops.
4. 04Structured outputs, tool use and function calling13% of the exam

   • Design reliable tool-calling and structured output interfaces between LLMs and software systems.
   • Scope: JSON Schema constrained outputs, function and tool definitions, Model Context Protocol (MCP), parameter validation, error handling and retries, deterministic parsing.
5. 05Context engineering and retrieval-augmented generation9% of the exam

   • Architect context and retrieval pipelines that maximise grounded, accurate responses within token budgets.
   • Scope: Chunking and embedding strategies, hybrid search and reranking, context assembly and compression, citation and grounding, long-context vs RAG trade-offs, prompt caching.
6. 06Agentic systems and multi-agent orchestration5% of the exam

   • Design agentic workflows with appropriate autonomy, control points and failure handling.
   • Scope: Agent loops, planning and memory, subagent delegation, orchestration patterns, guardrails on autonomy and tool permissions, human-in-the-loop checkpoints, failure recovery.
7. 07Evaluation engineering for LLM applications11% of the exam

   • Design evaluation frameworks that measure prompt and system quality and detect regressions reliably.
   • Scope: Golden datasets, LLM-as-judge and rubric design, pairwise comparison, regression suites, eval harnesses (promptfoo, OpenAI Evals, Inspect), hallucination and faithfulness metrics, statistical significance.
8. 08Prompt security: injection, jailbreaks and data leakage13% of the exam

   • Evaluate LLM applications for prompt-level threats and design layered mitigations.
   • Scope: Direct and indirect prompt injection, OWASP Top 10 for LLM Applications, data exfiltration via tools, system prompt leakage, defence in depth, input/output filtering, privilege separation.
9. 09Optimisation for cost, latency and model selection6% of the exam

   • Optimise LLM solutions for quality, cost and latency and justify model and technique choices.
   • Scope: Model routing and cascades, prompt compression, caching, batching, small vs frontier model trade-offs, prompt optimisation tools (DSPy), fine-tuning vs prompting decision criteria.
10. 10Multimodal and domain-specific prompting5% of the exam

    • Design effective prompts for multimodal, code and multilingual tasks and evaluate their limitations.
    • Scope: Vision and document understanding prompts, audio and speech pipelines, code generation prompting, multilingual and low-resource language considerations (including Albanian), domain adaptation.
11. 11Responsible AI governance and LLMOps5% of the exam

    • Lead governance and operational practices that keep LLM applications compliant, monitored and accountable.
    • Scope: EU AI Act obligations, ISO/IEC 42001 AI management systems, NIST AI RMF, prompt and model version control, monitoring and drift, bias and safety review, audit trails.

Question types

Single choice 100%

Delivery

Online (remote proctored), Test centre, On-site (corporate)

Retakes

Up to three attempts within any 12-month period. A 14-day waiting period applies after a first unsuccessful attempt and 30 days after a second. Each retake requires a new booking and fee. Retakes draw a different randomised form from the question bank.

Provisional cut score — psychometric validation is required before live delivery. Blueprint v1.0.

1. 1

   Register

   Create a free account and start the programme.

2. 2

   Practice

   Online practice test (coming soon).

3. 3

   Training

   Optional

   Online course or classroom — or skip it.

4. 4

   Certification

   Independent exam and certification decision.

Training is optional and does not by itself confer certification. Find a training partner