Skip to content

Learn by shipping, not collecting tutorials.

A personalized path from software foundations to production AI, with time estimates, prerequisites, practice tasks and proof-of-skill projects.

Progress
0%
Topics
0/23
Remaining
175h
At your pace
25 wk

Personalize your plan

AI application developer: Build useful LLM products, RAG systems and AI features. The full roadmap contains about 175 focused learning hours, excluding projects.

Your AI engineering path

Select a node to open its lesson plan. Follow the connections from foundations to production.

Best next action

Python for product development

Recommended for ai application developer. Spend about 8 hours, then complete the practice task and save the proof.

Roadmap views
  • Completed
  • Recommended
  • To learn

23 topics shown

82%

Stage 01

Software foundations

Write, test and connect the code every AI product depends on.

0/4 complete0%

Stage 02

Machine learning essentials

Learn enough ML to evaluate models and make sound product decisions.

0/4 complete0%

Stage 03

LLM application development

Move from impressive demos to controlled, testable AI features.

0/5 complete0%

Stage 04

RAG, tools and agents

Connect models to knowledge and actions without losing control.

0/5 complete0%

Stage 05

Production AI engineering

Ship systems that remain secure, observable and affordable after launch.

0/5 complete0%

Scroll vertically to move through every stage. Drag the horizontal scrollbar on smaller screens.

Stage 010/4 complete

Software foundations

Write, test and connect the code every AI product depends on.

Stage 020/4 complete

Machine learning essentials

Learn enough ML to evaluate models and make sound product decisions.

Stage 030/5 complete

LLM application development

Move from impressive demos to controlled, testable AI features.

Stage 040/5 complete

RAG, tools and agents

Connect models to knowledge and actions without losing control.

Stage 050/5 complete

Production AI engineering

Ship systems that remain secure, observable and affordable after launch.

Python for product development

Topic plan · Software foundations

  • Beginner
  • 8 focused hours
  • software

Use functions, modules, types, exceptions, virtual environments and packages confidently.

You should be able to

  • Write small programs without copying a tutorial
  • Structure code across modules
  • Handle invalid input and runtime errors
Practice task
Build a CLI that accepts a topic and returns a formatted study plan.
Proof of completion
A tested repository another developer can run.

Skill coverage

  • Software0%
  • ML & data0%
  • LLM systems0%
  • Agents0%
  • Production0%

Five checkpoints that prove your ability

Do not wait until the end. Ship one project after each stage and improve earlier work as your skills grow.

  1. After stage 01

    AI-ready API client

    Build a command-line program that calls a public API, validates input, handles errors and saves structured results.

    • Readable Python package
    • Environment-based configuration
    • Tests for success and failure paths
    • Clear README with setup steps
  2. After stage 02

    Support ticket classifier

    Train a model that routes support tickets into categories and explain where it fails.

    • Documented dataset split
    • Baseline and improved model
    • Precision/recall comparison
    • Failure analysis with examples
  3. After stage 03

    Structured AI assistant

    Build a streaming assistant that produces validated structured output and handles unsafe or invalid requests.

    • Prompt contract and examples
    • Structured output validation
    • Streaming interface
    • Small regression evaluation set
  4. After stage 04

    Cited research assistant

    Build an assistant that searches private documents, cites evidence and uses one external tool safely.

    • Document ingestion pipeline
    • Retrieval evaluation set
    • Answers with source citations
    • Tool permissions and failure handling
  5. After stage 05

    Production AI capstone

    Turn one previous project into a deployed product with users, monitoring, evaluation and cost controls.

    • Deployed application and runbook
    • Quality and latency dashboard
    • Security threat model
    • Post-launch evaluation report

Test knowledge, then build something real.

Use quizzes to find weak concepts and the practical AI course for deeper guided exercises.