✦Sparkl
innovator · thinks

Reasoning & Agents

Build AI systems that think step-by-step, use tools, and remember context. Learn how large language models chain reasoning, plan tasks, and act as autonomous agents.

20 modules·Difficulty: ★★★☆☆· 5 Free
Start module 1

Modules

  • 1
    Free8 min
    What Is Chain-of-Thought?
    Discover how showing your reasoning steps helps AI solve harder problems 🧠
  • 2
    Free10 min
    Zero-Shot vs Few-Shot Reasoning
    Compare prompting strategies and learn when examples boost accuracy.
  • 3
    Free9 min
    Planning vs Reacting
    Explore the difference between agents that plan ahead and those that respond step-by-step.
  • 4
    Free11 min
    Self-Consistency Sampling
    Learn how generating multiple reasoning paths and voting improves reliability.
  • 5
    Free10 min
    Decomposing Complex Tasks
    Break big problems into sub-questions the AI can tackle one at a time.
  • 6
    Paid12 min
    Function Calling Basics
    Give your AI the power to call external functions and retrieve real data.
  • 7
    Paid11 min
    Tool Selection Strategies
    Teach agents to pick the right tool for each sub-task automatically.
  • 8
    Paid10 min
    Short-Term Memory Buffers
    Store recent conversation turns so your agent remembers what just happened.
  • 9
    Paid12 min
    Embedding-Based Retrieval
    Use vector embeddings to find relevant memories from a large knowledge base.
  • 10
    Paid11 min
    ReAct: Reasoning + Acting
    Combine thought traces with actions in a loop that refines answers iteratively.
  • 11
    Paid10 min
    Building a Task Planner
    Code an agent that outlines steps before executing them sequentially.
  • 12
    Paid12 min
    Error Handling & Retries
    Implement fallback logic so agents recover gracefully when tools fail.
  • 13
    Paid11 min
    Prompt Templates & Variables
    Design reusable prompt skeletons with placeholders for dynamic inputs.
  • 14
    Paid10 min
    Multi-Agent Collaboration
    Orchestrate multiple specialized agents that divide and conquer complex workflows.
  • 15
    Paid12 min
    Evaluating Agent Output
    Create test suites with assertions to measure correctness and consistency.
  • 16
    Paid11 min
    Document Parsing & Chunking
    Split large texts into manageable pieces for efficient retrieval and reasoning.
  • 17
    Paid12 min
    Citation & Source Tracking
    Ensure your agent cites where it found each fact to build trust and transparency.
  • 18
    Paid10 min
    Query Refinement Loop
    Teach agents to rephrase searches when initial results are incomplete or off-topic.
  • 19
    Paid12 min
    Capstone: Research Agent
    Combine planning, tools, memory, and citations into one autonomous research assistant.
  • 20
    Final exam20 min
    Final Exam: Reasoning & Agents
    Prove your mastery of chains, tools, memory, and multi-step agent design.