AI Engineering Course - Kotlin Implementation
Fredrik Rødland's Kotlin implementations of the Scrimba AI Engineering Path course assignments and examples. While the original course focuses on JavaScript, I chose Kotlin to concentrate on learning AI engineering concepts without syntax distractions. Built on Spring Boot with extensive use of Server-Sent Events (SSE) for real-time communication.
Throughout development, I've relied heavily on Claude Code CLI as both a development accelerator and learning tool. This repository serves as a personal reference and demonstrates how traditional backend technologies combine with cutting-edge AI capabilities.
Built: 2025-06-19 08:27:30 - Git: fac6718
Core AI
Prompt Lab ✓ Done
Compare different prompting techniques to see how small changes dramatically impact AI response quality and accuracy. Essential foundation for all AI interactions.
View Demo
Chat Basics
✓ Done
Interactive form to generate AI responses with customizable prompts, system messages, and model parameters.
Foundation demo from the Scrimba course.
Chat Studio ✓ Done
Full conversation experience with context awareness, session management, and persistent history. Build ChatGPT-like interactions with memory.
View DemoTemperature Play ✓ Done
Discover how temperature affects AI output randomness. Compare creative vs consistent responses with the same prompt using different temperature values.
View DemoTop-P Magic ✓ Done
Explore how top-p affects AI vocabulary selection and word diversity.
See how nucleus sampling shapes response quality and variety in real-time.
Fine-Tune Factory Maybe Next Year
Train specialized models on your data using OpenAI's fine-tuning API.
Create domain-specific AI models for improved task performance.
Vector Space
✓ Done
Convert text into high-dimensional vectors using OpenAI's models.
Essential foundation for semantic search with detailed analysis and visualization.
Text Slicer ✓ Done
Interactive chunking tool with customizable size and overlap settings.
Essential for preparing large texts for AI processing and RAG systems.
PolyGlot
✓ Done
Solo Project
Translate between dozens of languages while maintaining conversation flow and cultural nuances.
Node.js/JavaScript app with Socket.io real-time chat.
Data
Stock Oracle ✓ Done
Fetch real-time stock data from Polygon.io and generate intelligent market analysis.
Complete pipeline from data retrieval to AI-driven insights.
DodgyDave
✓ Done
Solo Project
Give the app a couple of tickers and it'll tell you what to do - buy or sell!
Vite/JavaScript app with Socket.io real-time chat.
Vector Vault ✓ Done
Explore semantic search using embeddings and PostgreSQL pgvector extension.
Foundation for RAG systems using Supabase as the vector store.
RAG Engine ✓ Done
Advanced RAG implementation from the Scrimba course.
Complete pipeline: query→embedding→vector search→context-aware AI response using movie/podcast database.
PopChoice Dreams
Solo ProjectAnalyze preferences, mood, and viewing history to find the perfect film for any occasion.
Intelligent recommendations powered by preference learning.
Open Source
Text Classifier
✓ Done
Classify text into custom categories using HuggingFace BART model.
Explore zero-shot classification with confidence scores and detailed analysis.
Text Squisher
✓ Done
Compress long texts into concise summaries using HuggingFace BART model.
Customizable length parameters for perfect summary control.
Object Spotter
✓ Done
Detect and locate objects in images using DETR (Detection Transformer) model.
HuggingFace-powered vision with bounding boxes and confidence scores.
Vision & Image
Poster Lab ✓ Done
Combine cinematic imagery with artistic styles using advanced image generation.
Fine-tune parameters, models, and quality settings for stunning poster results.
Image Remix ✓ Done
Edit existing images with DALL-E 2 by describing what you want to change.
Precise mask-based editing for specific areas and seamless modifications.
Vision Quest ✓ Done
Analyze images with GPT-4 Vision using our curated gallery of demo images.
Explore AI's visual understanding with suggested prompts and detailed insights.
AI Agents
ReAct Playground ✓ Done
AI agent that solves problems through iterative thinking and action.
Watch step-by-step reasoning as the agent plans, acts, and learns from results.
Tool Master ✓ Done
AI directly calls external tools using structured function schemas.
Compare OpenAI's native approach with ReAct's manual reasoning patterns.
Assistant Builder ✓ Done
Build persistent AI assistants using OpenAI's Assistants API with file handling, code execution, and long-term memory.
Advanced stateful assistants with tool capabilities beyond simple chat interactions.
Trip Wizard ✓ Done
Multi-agent system using parallel processing to plan comprehensive trips.
Research, weather, activities, and food agents working together in real-time.
MCP Protocol
MCP Toolbox ✓ Done
Expose tools and resources that Claude Desktop can discover and use automatically.
Create your own AI tool ecosystem using Spring AI and MCP protocol.
MCP Decoded ✓ Done
Educational demo showing JSON-RPC 2.0 calls, tool discovery, and parameter schemas.
Learn Model Context Protocol mechanics step-by-step.
MCP Wizard ✓ Done
AI assistant that dynamically discovers and uses external tools via MCP.
Chat interface with OpenAI function calling and persistent session memory.
Frameworks
LangChain Lab
✓ Done
Test text analysis, summarization, and sentiment classification using annotated interfaces.
Compare LangChain4j's approach with our custom implementations.
Spring AI
✓ Done
Spring's AI integration with auto-configuration, dependency injection, and native patterns.
Compare against LangChain4j and custom implementations.
OpenAI Toolkit Procrastinating
Type-safe builders, streaming support, and structured outputs using OpenAI's official library.
Compare with our custom HTTP implementation approach.