1. Agentic AI

- Has built and deployed AI agents, not just chatbots, agents that call tools/APIs and complete multi-step tasks

- Hands-on with agent frameworks: Lang Graph, Lang Chain, Semantic Kernel, Auto Gen or CrewAI

- Tool/function calling, MCP (Model Context Protocol), multi-agent workflows

- Human-in-the-loop: agent pauses for human approval before risky actions

- Agent state and memory (e.g. Redis, Postgres), guardrails for safety

2. Generative AI

- 2.5–3+ years hands-on GenAI work (not just general AI/ML)

- LLM APIs: Azure OpenAI / OpenAI

- RAG, embeddings, vector databases (Azure AI Search, pgvector, Pinecone, FAISS)

- Prompt engineering and structured outputs

3. Engineering and production

- Robust Python, FastAPI, Docker, Kubernetes/AKS, CI/CD

- Testing and evaluation of AI outputs (golden test sets, Ragas, Deep Eval, LLM-as-judge)

- Monitoring and observability (Open Telemetry, Lang Smith, Langfuse, App Insights/Dynatrace)

- Cost, reliability and PII/security handling

4. Platform mindset

- Built reusable components, templates, SDKs or frameworks used by multiple teams

📌 AI Engineer (India)
🏢 LTM
📍 India