AI Engineer Roadmap

With GenAI approaching at a rapid rate and everyone trying to find the most optimal ways to implement AI into their products and workflows, theres never been a better time to learn AI Engineering.

What is AI Engineering?

AI Engineering is the process of designing and implementing AI systems using pre-trained models and existing AI tools to solve practical problems. AI Engineers focus on applying AI in real-world scenarios, improving user experiences, and automating tasks, without developing new models from scratch. They work to ensure AI systems are efficient, scalable, and can be seamlessly integrated into business applications, distinguishing their role from AI Researchers and ML Engineers, who concentrate more on creating new models or advancing AI theory.

AI Engineer vs ML Engineer

An AI Engineer uses pre-trained models and existing AI tools to improve user experiences. They focus on applying AI in practical ways, without building models from scratch. This is different from AI Researchers and ML Engineers, who focus more on creating new models or developing AI theory.

AI Engineering Roadmap

The following is a list of popular topics related to AI Engineering, for a more in depth and interactive roadmap, visit the official AI Engineering roadmap.

Common Terms

LLMs
Inference
Training
Embeddings
Vector Databases
RAG
Prompt Engineering
AI Agents
AI vs AGI

Using Pre-trained Models

Pre-trained Models
Benefits of Pre-trained Models
Limitations and Considerations
Popular AI Models

Open AI Models

Capabilities / Context Length
Cut-off Dates / Knowledge

Other Popular Models

Anthropic's Claude
Google's Gemini
Azure AI
AWS Sagemaker
Hugging Face Models
Mistral AI
Cohere

OpenAI Platform

OpenAI API
Chat Completions API
Writing Prompts
OpenAI Playground
Fine-tuning
Managing Tokens
Maximum Tokens
Token Counting
Pricing Considerations

AI Safety and Ethics

Understanding AI Safety Issues
Prompt Injection Attacks
Bias and Fairness
Security and Privacy Concerns
Conducting Adversarial Testing
OpenAI Moderation API
Adding End-user IDs in Prompts
Robust Prompt Engineering
Know your Customers / Use Cases
Constraining Outputs and Inputs
Safety Best Practices

Open Source AI

Open vs Closed Source Models
Popular Open Source Models
Hugging Face
- Hugging Face Hub
- Hugging Face Tasks
- Finding Open Source Models
- Using Open Source Models
Inference SDK
Transformers.js
Ollama
- Ollama Models
- Ollama SDK

Embeddings

Semantic Search
Recommendation Systems
Anomaly Detection
Data Classification

Embeddings & Vector Databases

Use Cases for Embeddings
OpenAI Embeddings API
OpenAI Embedding Models
Pricing Considerations
Open-Source Embeddings
- Sentence Transformers
- Models on Hugging Face

Vector Databases

Purpose and Functionality
Popular Vector DBs
- Chroma
- Pinecone
- Weaviate
- FAISS
- LanceDB
- Qdrant
- Supabase
- MongoDB Atlas
Indexing Embeddings
Performing Similarity Search
Implementing Vector Search

RAG & Implementation

RAG Use Cases
RAG vs Fine-tuning
Chunking
Embedding
Vector Database
Retrieval Process
Generation
Implementing RAG
- Ways of Implementing RAG
- Using SDKs Directly
- LangChain
- Llama Index
- OpenAI Assistant API
- Replicate

Prompt Engineering

ReAct Prompting
Manual Implementation
OpenAI Functions / Tools
OpenAI Assistant API
Building AI Agents

Multimodal AI

Multimodal AI Use Cases
- Image Understanding
- Image Generation
- Video Understanding
- Audio Processing
- Text-to-Speech
- Speech-to-Text
Multimodal AI Tasks
- OpenAI Vision API
- DALL-E API
- Whisper API
- Hugging Face Models
- LangChain for Multimodal Apps
- LlamaIndex for Multimodal Apps
- Implementing Multimodal AI

Next Steps

At roadmap.sh we have the comprehensive AI Engineering roadmap as well as the following that can help you on your travels!

Blog

Dan