AI Agents

AI Agents

Learn to design, build and ship AI agents in @currentYear@

This page is a visitor-friendly snapshot of how William currently sees this area: what feels strong, what is still being refined, and where the evidence is still partial.

0

Done

0

Learning

0

Target

0

Skipped

101

Tracked

Status meanings

Done: enough evidence to present this area as a solid strength.

Learning: an area of active development with growing confidence.

Target: relevant and intentional, but still earlier in the journey.

Skipped: not a current priority relative to other areas.

How the estimates are formed

Profile and manual review: 0

Evidence from project experience: 0

Competency mapping alignment: 0

Quiz-backed updates: 0

These signals help estimate progress, but they are still best treated as directional rather than perfectly complete.

Areas currently being strengthened

No standout growth areas are being highlighted here yet.

Basic Backend Development

Untracked

Git and Terminal Usage

Untracked

REST API Knowledge

Untracked

Transformer Models and LLMs

Untracked

Tokenization

Untracked

Context Windows

Untracked

Token Based Pricing

Untracked

Temperature

Untracked

Frequency Penalty

Untracked

Presence Penalty

Untracked

Top-p

Untracked

Stopping Criteria

Untracked

Open Weight Models

Untracked

Closed Weight Models

Untracked

Streamed vs Unstreamed Responses

Untracked

Reasoning vs Standard Models

Untracked

Fine-tuning vs Prompt Engineering

Untracked

Embeddings and Vector Search

Untracked

Understand the Basics of RAG

Untracked

Pricing of Common Models

Untracked

What are AI Agents?

Untracked

What are Tools?

Untracked

Agent Loop

Untracked

Observation & Reflection

Untracked

Perception / User Input

Untracked

Reason and Plan

Untracked

Acting / Tool Invocation

Untracked

Personal assistant

Untracked

Code generation

Untracked

Data analysis

Untracked

Web Scraping / Crawling

Untracked

NPC / Game AI

Untracked

Max Length

Untracked

What is Prompt Engineering

Untracked

Be specific in what you want

Untracked

Provide additional context

Untracked

Use relevant technical terms

Untracked

Use Examples in your Prompt

Untracked

Iterate and Test your Prompts

Untracked

Specify Length, format etc

Untracked

Tool Definition

Untracked

Web Search

Untracked

Code Execution / REPL

Untracked

Database Queries

Untracked

API Requests

Untracked

Email / Slack / SMS

Untracked

File System Access

Untracked

What is Agent Memory?

Untracked

Short Term Memory

Untracked

Long Term Memory

Untracked

Episodic vs Semantic Memory

Untracked

RAG and Vector Databases

Untracked

User Profile Storage

Untracked

Summarization / Compression

Untracked

Forgetting / Aging Strategies

Untracked

ReAct (Reason + Act)

Untracked

Model Context Protocol (MCP)

Untracked

MCP Hosts

Untracked

MCP Client

Untracked

MCP Servers

Untracked

Creating MCP Servers

Untracked

Local Desktop

Untracked

Remote / Cloud

Untracked

Chain of Thought (CoT)

Untracked

RAG Agent

Untracked

Planner Executor

Untracked

DAG Agents

Untracked

Tree-of-Thought

Untracked

Manual (from scratch)

Untracked

LLM Native "Function Calling"

Untracked

OpenAI Functions Calling

Untracked

Gemini Function Calling

Untracked

OpenAI Assistant API

Untracked

LangChain

Untracked

LlamaIndex

Untracked

Haystack

Untracked

AutoGen

Untracked

CrewAI

Untracked

Smol Depot

Untracked

Anthropic Tool Use

Untracked

Metrics to Track

Untracked

Unit Testing for Individual Tools

Untracked

Integration Testing for Flows

Untracked

Human in the Loop Evaluation

Untracked

LangSmith

Untracked

Ragas

Untracked

DeepEval

Untracked

Structured logging & tracing

Untracked

LangSmith

Untracked

Helicone

Untracked

LangFuse

Untracked

openllmetry

Untracked

Prompt Injection / Jailbreaks

Untracked

Tool sandboxing / Permissioning

Untracked

Data Privacy + PII Redaction

Untracked

Bias & Toxicity Guardrails

Untracked

Safety + Red Team Testing

Untracked

Agno

Untracked

LangGraph

Untracked

Self-critique Agents

Untracked

Multi-Agents

Untracked