Technology Stacks in Generative AI

"Generative AI acts as the creative engine within Natural Language Processing (NLP), enabling machines to generate human-quality text, translate languages, and create new content based on the patterns it learns from vast amounts of text data."- Gemini 2025

Generative AI Stack - Skills in Demand

A snapshot of technologies and products seen in recent job postings

This snapshot is from a sampling of job postings taken in April 2024. The landscape for Generative AI is evolving rapidly, new technologies and techniques emerge continuously across the stack and, like the big bang , products are popping up everywhere.

* Added by Claude AI in June 2025

  • Programming Languages
    • Python
    • * JavaScript/TypeScript - Major growth in web-based AI applications
    • * Rust - Increasingly popular for high-performance AI infrastructure
    • * Go - Growing adoption for AI microservices and tooling
  • ML Frameworks
    • Tensorflow
    • Pytorch
    • * JAX - Google's framework gaining traction for research and production
    • * Hugging Face Transformers - Essential library that's become standard
    • * vLLM - High-performance inference serving framework
  • Generative AI Frameworks
    • Langchain
    • Llama Index
    • OpenAI Eval
    • Ollama
    • * LangSmith - LangChain's production monitoring/debugging platform
    • * CrewAI - Multi-agent framework that's gained popularity
    • * AutoGen - Microsoft's multi-agent conversation framework
    • * DSPy - Stanford's framework for optimizing LM prompts/weights
    • * Haystack - Production-ready NLP framework
    • * LiteLLM - Unified API for 100+ LLMs
  • Vector Embeddings
    • Sentence Transformers
    • Universal Sentence Encoders
    • OpenAI Embeddings
    • * Cohere Embeddings - Strong competitor to OpenAI
    • * BGE Models - BAAI's high-performing embedding models
    • * E5 Models - Microsoft's multilingual embeddings
    • * Voyage AI Embeddings - Specialized for retrieval tasks
  • Vector Database
    • Pinecone
    • Weaviate
    • Milvus
    • pgvector
    • Doc-Array (might be less relevant now)
    • Chroma
    • * Qdrant - Growing rapidly in popularity
    • * Zilliz Cloud - Managed Milvus service
    • * Redis Vector Search - Redis with vector capabilities
    • * SingleStore - Distributed SQL with vector search
  • Deployment Platforms
    • AWS
    • GCP
    • Render
    • * Azure - Major player, especially with OpenAI partnership
    • * Hugging Face Spaces - Popular for ML model demos
    • * Replicate - AI model hosting platform
    • * Modal - Serverless cloud for AI/ML
    • * RunPod - GPU cloud platform
    • * Together AI - Inference platform
  • * LLM APIs & Models
    • OpenAI (GPT-4, GPT-4o, o1)
    • Anthropic Claude
    • Google Gemini
    • Meta Llama
    • Cohere
    • Mistral AI
    • Together AI
    • Groq (for inference)
  • * AI Development Platforms
    • Hugging Face Hub
    • Weights & Biases
    • MLflow
    • Vertex AI
    • Azure OpenAI Service
    • Amazon Bedrock
  • * Observability & Monitoring
    • LangSmith
    • Weights & Biases
    • Phoenix (Arize)
    • Helicone
    • LangFuse
  • * Fine-tuning & Training
    • Unsloth
    • LoRA/QLoRA
    • Axolotl
    • TRL (Transformer Reinforcement Learning)