- Design and develop single- and multi-agent systems using frameworks like Lang Graph, Lang Chain, CrewAI, AutoGen, or similar.

- Build agent capabilities including tool use, planning, memory, reflection, and human-in-the-loop workflows.

- Integrate agents with external systems via APIs, MCP (Model Context Protocol) servers, and custom connectors.

- Implement RAG pipelines using vector databases (Pinecone, Weaviate, Qdrant, Chroma, etc.) for context-aware agent behavior.

- Optimize prompts, agent loops, and orchestration for accuracy, latency, and cost.

- Build evaluation harnesses, guardrails, and observability for agent performance (LangSmith, Langfuse, Arize, or similar).

- Collaborate with product, data, and engineering teams to identify high-value agentic use cases and ship them to production.

- Stay current with advances in LLMs, reasoning techniques, and agent architecture.

Key skills and experience required:

- Strong Python programming skills, including async patterns.

- Hands-on experience building LLM-powered applications with frameworks such as LangChain, LangGraph, LlamaIndex, CrewAI, or AutoGen.

- Solid understanding of agent design patterns: ReAct, planning, tool use, multi-agent collaboration, and memory.

- Experience with leading LLM APIs (Anthropic Claude, OpenAI, Google Gemini, open-source models).

- Working knowledge of RAG, embeddings, and vector databases.

- Prompt engineering and structured output techniques (function calling, JSON schemas, tool definitions).

- Experience deploying production systems on AWS / Azure / GCP, with Docker and CI/CD.

- Experience with MCP (Model Context Protocol), browser-use, or computer-use agents.

- Fine-tuning, LoRA, or model distillation experience.

- Knowledge of evaluation frameworks (RAGAS, DeepEval, custom evals).

- Experience with workflow orchestration (Temporal, Airflow, Prefect).

- Familiarity with safety, alignment, and red-teaming practices for LLM systems.

- Background in NLP, ML, or applied research.