Super AI

Super AI Technologies

The technologies behind modern enterprise AI — what each one is, how it works, the tools involved, and the skills employers are hiring for.

10 topics

01

AI Agents & Agent Development

Building models that plan, use tools and complete multi-step work — with the guardrails, oversight and evaluation production needs.

  • Tool calling
  • MCP
  • Orchestration
  • Human-in-the-loop
 02

AI Security, Safety & Governance

Prompt injection, data leakage, guardrails and red teaming — plus the frameworks (NIST AI RMF, ISO/IEC 42001, EU AI Act) that govern AI use.

  • OWASP LLM Top 10
  • Guardrails
  • Red teaming
  • Compliance
 03

Embeddings & Vector Search

Turning content into vectors for semantic search, recommendations, deduplication and clustering — and storing them at scale.

  • Embedding models
  • ANN indexes
  • Vector databases
  • Similarity
 04

Evaluation & LLMOps

Evaluation sets, LLM-as-judge, tracing, cost monitoring and release discipline for AI systems that must stay good after launch.

  • Eval sets
  • LLM-as-judge
  • Tracing
  • Cost monitoring
 05

Fine-Tuning & Model Customization

Supervised and preference tuning, LoRA adapters, distillation and quantization — and the decision of when tuning beats prompting or RAG.

  • LoRA / QLoRA
  • Preference tuning
  • Distillation
  • Quantization
 06

Generative AI

Text, image, audio, video and code generation, plus multimodal understanding — with the brand, IP and provenance controls enterprises need.

  • Multimodal
  • Diffusion
  • Speech
  • Content provenance
 07

Large Language Models (LLMs)

How transformer-based language models generate text, reason and call tools — and how to pick, host and budget for them.

  • Transformers
  • Tokens & context
  • Model selection
  • Inference
 08

Machine Learning & Deep Learning Foundations

Supervised and unsupervised learning, model evaluation, gradient boosting, neural networks and MLOps — the fundamentals behind every AI role.

  • Supervised learning
  • Neural networks
  • Metrics
  • MLOps
 09

Prompt & Context Engineering

Writing instructions, examples and schemas — and curating exactly what goes into the context window — for reliable model behavior.

  • System prompts
  • Few-shot
  • Structured output
  • Context curation
 10

Retrieval-Augmented Generation (RAG)

Retrieve the right passages from your knowledge base at question time and have the model answer from them, with citations.

  • Chunking
  • Hybrid search
  • Reranking
  • Citations

AI Learning

See the core technologies working together on one realistic back-office process.

AI

AI Learning

A complete invoicing system built end to end: architecture decisions, pipeline DAG, AI-driven lifecycle, enterprise MCP gateway and RAG over contracts.

  • ADRs
  • DAGs
  • AI-DLC
  • Enterprise MCP
  • RAG

How the pieces fit together

Real AI systems combine several of these technologies. A typical enterprise assistant uses an LLM, steered by carefully engineered context, grounded by RAG over a vector index, able to act through agent tools, and kept trustworthy by evaluation, monitoring and governance.

  1. Foundations

    Machine learning and deep learning fundamentals: data, training, validation and metrics.

  2. Models

    Large language and generative models, chosen and hosted for quality, cost and control.

  3. Instructions

    Prompt and context engineering that makes model behavior predictable.

  4. Knowledge

    Embeddings, vector search and RAG that ground answers in your own data.

  5. Action

    Agents that use tools and complete multi-step work with humans in control.

  6. Production

    Fine-tuning, evaluation, LLMOps, security and governance that keep it all reliable.

Put these technologies to work

MrAI.dev helps enterprises choose, build and operate AI systems across this whole stack — from a first use case to production governance.

Book a discovery call