Skip to main content

Learning Roadmap

How to Become a Prompt Systems Designer

A step-by-step, phase-based learning path from beginner to job-ready Prompt Systems Designer. Estimated completion: 6 months across 4 phases.

4 Phases
24 Weeks Total
Medium Entry Barrier
Advanced Difficulty
Your Progress 0 / 4 phases

Progress saved in your browser — no account needed.

  1. Foundations: LLMs & Basic Prompting

    4 weeks
    • Understand core transformer concepts and LLM text generation
    • Master basic prompt patterns (zero-shot, few-shot, instruction-following)
    • Learn to use the OpenAI and Anthropic APIs for simple tasks
    • Anthropic's 'Introduction to Prompt Engineering' documentation
    • OpenAI's 'GPT Best Practices' guide
    • Fast.ai 'Practical Deep Learning for Coders' (NLP sections)
    Milestone

    Can design and implement effective basic prompts for classification, summarization, and Q&A tasks via API.

  2. System Engineering: Chains & RAG

    6 weeks
    • Learn to design and build prompt chains using LangChain
    • Implement a basic Retrieval-Augmented Generation (RAG) system
    • Understand structured data output (JSON mode, function calling)
    • LangChain documentation and tutorial series
    • DeepLearning.AI 'LangChain for LLM Application Development' course
    • Pinecone or Weaviate 'RAG 101' learning resources
    Milestone

    Can architect and prototype a multi-step RAG system that answers questions from a custom knowledge base.

  3. Advanced Systems: Agents, Evaluation & Safety

    8 weeks
    • Design agent systems with tool use and planning capabilities
    • Build rigorous evaluation frameworks for prompt systems
    • Implement safety guardrails and content moderation layers
    • DeepLearning.AI 'AI Agents in LangGraph' course
    • LangSmith documentation for tracing and evaluation
    • Research papers on LLM evaluation (e.g., 'Chatbot Arena') and safety
    Milestone

    Can design an agent with custom tools, write comprehensive eval suites, and deploy guardrails to block harmful outputs.

  4. Productionization & Specialization

    6 weeks
    • Learn CI/CD for prompt systems and versioning strategies
    • Explore advanced optimization (DSPy, prompt tuning)
    • Develop domain-specific expertise (e.g., legal, coding, healthcare)
    • DSPy documentation for optimizing LM pipelines
    • MLOps resources for prompt management at scale
    • Case studies from companies like Stripe, Duolingo, or Morgan Stanley on LLM integration
    Milestone

    Can manage prompt systems as production-grade software, optimize them programmatically, and apply deep expertise to a vertical domain.

Practice Projects

Apply your skills with hands-on projects. Ordered by difficulty.

Multi-Modal Research Assistant

Advanced

Build a system that accepts a PDF research paper (as text) and a voice query (transcribed). It should summarize the paper, answer questions about it, and extract key figures/tables, all while citing specific sections.

~30h
RAG ArchitectureMulti-Modal PromptingCitation & Attribution Design

Resilient Customer Support Agent

Intermediate

Create an agent that handles support tickets for a fictional e-commerce site. It must follow a strict policy, handle angry users, escalate appropriately, and resist basic prompt injection attempts.

~25h
System Prompt DesignSafety GuardrailsPersona Development

Prompt-as-Code DevOps Pipeline

Intermediate

Set up a Git repository where prompts are versioned YAML/JSON files. Create a GitHub Action that, on commit, runs an evaluation suite (using a small dataset) and posts the results as a PR comment.

~20h
Prompt Version ControlCI/CD for AIEvaluation Frameworks

Domain-Specific Code Review Bot

Advanced

Design a prompt system that takes a code diff and provides insightful review comments based on a given coding style guide (e.g., Google's Python style). It should explain why changes are suggested.

~35h
Structured Data OutputFew-Shot with Code ExamplesContext Window Management

Interactive Story Generator with Consistent Memory

Beginner

Build a text-based game where the AI narrator remembers character names, plot points, and world details from previous interactions, maintaining consistency over a long session.

~15h
Long-Context PromptingState Tracking via PromptingPersona & World Building

Ready to Start Your Journey?

Prep for interviews alongside your learning — it reinforces every concept.