Sponsor slotYour AI dev tool here · Reach engineering leaders tracking AI adoption.
EngAI Signal

Tags

Every AI tool, model, framework, practice and governance topic we detect in engineering job postings. Counts are open roles mentioning the tag, across all engineering roles; leadership roles in brackets. Tags with fewer than 5 open roles are not listed.

Tools

  1. Claude Codecounts as AI codingAnthropic's agentic coding tool.441open roles (87 leadership)
  2. GitHub Copilotcounts as AI codingGitHub's AI pair programmer. Microsoft 365, Studio and Sales copilots are excluded.336open roles (60 leadership)
  3. Cursorcounts as AI codingAI code editor from Anysphere.323open roles (64 leadership)
  4. OpenAI Codexcounts as AI codingOpenAI's coding agent.156open roles (21 leadership)
  5. OpenCodecounts as AI codingOpen-source AI coding agent for the terminal.23open roles (2 leadership)
  6. Windsurfcounts as AI codingAI IDE, formerly Codeium.22open roles (3 leadership)
  7. Gemini Code Assist / Gemini CLIcounts as AI codingGoogle's AI coding assistant and terminal agent.18open roles (5 leadership)
  8. JetBrains AI / Juniecounts as AI codingJetBrains AI Assistant and the Junie coding agent.8open roles (0 leadership)
  9. Aidercounts as AI codingOpen-source AI pair programmer in the terminal.5open roles (0 leadership)

Models & platforms

  1. ClaudeAnthropic's Claude models. Claude Code is counted separately, under Tools.348open roles (64 leadership)
  2. OpenAIOpenAI's models and API. Azure OpenAI and Codex are counted separately.285open roles (56 leadership)
  3. Open / local modelsOpen-weight and self-hosted language models: vLLM, Mistral, Llama, Ollama and others.173open roles (22 leadership)
  4. vLLMOpen-source inference server for language models.59open roles (12 leadership)
  5. Local / open-weight LLMsOpen-weight models and language models run locally, on-prem or self-hosted.57open roles (6 leadership)
  6. MistralMistral AI's models, including Mixtral and Codestral.48open roles (2 leadership)
  7. LlamaMeta's Llama models, including Code Llama. LlamaIndex is a separate tag.17open roles (2 leadership)
  8. OllamaTool for running language models locally.8open roles (0 leadership)
  9. QwenAlibaba's open-weight models.8open roles (2 leadership)
  10. DeepSeekDeepSeek's open-weight models.8open roles (2 leadership)
  11. AnthropicAnthropic as a model provider.154open roles (22 leadership)
  12. GeminiGoogle's Gemini models and API. Gemini Code Assist and Gemini CLI are counted under Tools.135open roles (19 leadership)
  13. ChatGPT / GPT modelsChatGPT and OpenAI's GPT model family.96open roles (21 leadership)
  14. Vertex AIGoogle Cloud's AI platform.93open roles (11 leadership)
  15. Amazon BedrockAWS's managed service for foundation models.90open roles (19 leadership)
  16. Hugging FaceThe model hub and its libraries.79open roles (7 leadership)
  17. Azure OpenAIOpenAI models hosted on Azure, including Azure AI Foundry.56open roles (2 leadership)
  18. Amazon SageMakerAWS's platform for training and serving ML models.48open roles (5 leadership)

Frameworks

  1. LangChainFramework for building LLM applications.211open roles (29 leadership)
  2. LangGraphLangChain's framework for stateful, multi-step agents.178open roles (32 leadership)
  3. LlamaIndexFramework for connecting LLMs to data.86open roles (7 leadership)
  4. LLM observabilityTracing and monitoring LLM applications: LangSmith, Langfuse and similar tools.76open roles (9 leadership)
  5. CrewAIFramework for teams of cooperating agents.45open roles (15 leadership)
  6. AutoGen / AG2Microsoft's multi-agent framework and its AG2 fork.33open roles (7 leadership)
  7. Semantic KernelMicrosoft's SDK for LLM applications and agents.14open roles (1 leadership)
  8. Pydantic AIAgent framework from the Pydantic team.14open roles (2 leadership)
  9. LiteLLMGateway and SDK for calling many model providers through one API.13open roles (2 leadership)
  10. DSPyFramework for programming, rather than prompting, language models.10open roles (1 leadership)
  11. Vercel AI SDKTypeScript toolkit for LLM applications.8open roles (2 leadership)

Practices

  1. Agentic workflowsAI agents that plan and act across tools and steps.1,267open roles (370 leadership)
  2. AI-assisted developmentcounts as AI codingGeneric mentions of AI-assisted coding and AI coding assistants.922open roles (213 leadership)
  3. RAGRetrieval-augmented generation.509open roles (116 leadership)
  4. Agentic codingcounts as AI codingCoding with autonomous or semi-autonomous coding agents.434open roles (116 leadership)
  5. MCP (Model Context Protocol)The open protocol for connecting AI models to tools and data.415open roles (102 leadership)
  6. Vector databasesVector stores and vector search for retrieval: pgvector, Pinecone, Weaviate, Qdrant and others.351open roles (79 leadership)
  7. pgvectorVector search extension for PostgreSQL.38open roles (3 leadership)
  8. PineconeManaged vector database.33open roles (3 leadership)
  9. WeaviateOpen-source vector database.23open roles (3 leadership)
  10. QdrantOpen-source vector database.20open roles (2 leadership)
  11. MilvusOpen-source vector database (Zilliz).16open roles (3 leadership)
  12. ChromaOpen-source embedding database.13open roles (2 leadership)
  13. FAISSMeta's library for vector similarity search.6open roles (1 leadership)
  14. Multi-agent systemsSeveral cooperating agents, and frameworks for orchestrating them.348open roles (84 leadership)
  15. Evals / LLM evaluationEvaluating LLM and agent behaviour.309open roles (79 leadership)
  16. Prompt / context engineeringDesigning prompts and the context given to models and agents.272open roles (71 leadership)
  17. Fine-tuning / LoRA / PEFTAdapting models with further training, including LoRA and other parameter-efficient methods.236open roles (52 leadership)
  18. EmbeddingsVector embeddings for search, retrieval and similarity. Counted only in an AI or vector-search sentence.138open roles (24 leadership)
  19. AI / agent harnessInfrastructure that runs, constrains and evaluates AI agents.131open roles (48 leadership)
  20. Spec-driven developmentcounts as AI codingWriting specs that AI agents implement (SDD).37open roles (4 leadership)
  21. AI code reviewcounts as AI codingAI-assisted or automated code review.18open roles (10 leadership)

Governance

  1. LLMOps / MLOpsOperating ML and LLM systems in production.415open roles (51 leadership)
  2. GuardrailsGuardrails on model and agent behaviour.320open roles (117 leadership)
  3. AI securitySecuring AI systems: prompt injection, red teaming, model security.128open roles (28 leadership)
  4. Responsible AIResponsible, trustworthy and ethical AI.122open roles (31 leadership)
  5. AI governanceGovernance and risk management of AI systems.67open roles (20 leadership)
  6. EU AI ActThe EU regulation on artificial intelligence.48open roles (9 leadership)