• Design and improve LLM-powered agents for task planning, execution, retrieval, and memory.
• Build modular agent stacks using frameworks like LangChain, CrewAI, AutoGen, or custom orchestrators.
• Develop workflows that connect agents to APIs, computer vision models, and vector databases.
• Lead prompt engineering, RAG pipelines, and multi-agent coordination strategies.
• Collaborate with engineers to productionize agents with proper observability.
• Evaluate agent behavior through logging, telemetry, and custom evaluation tools.
• Contribute to architecture patterns such as task workflows, tool interfaces, memory optimization, and feedback handling.
• Work with infrastructure teams to deploy agentic systems to production, ensuring monitoring, scaling, and rollback safety.

Must-Have Skills

• 2+ years building LLM-driven applications.
• Strong Python skills with experience in reusable agent stacks or RAG pipelines.
• Hands-on with LLM APIs (OpenAI, Anthropic, HuggingFace, etc.) and building prompt/tool workflows.
• Experience integrating vector databases (Weaviate, FAISS, Qdrant, Pinecone, Elasticsearch).
• Familiarity with multi-agent orchestration (CrewAI, AutoGen, LangGraph, or custom).
• Experience shipping AI tools or services into production with containerization and cloud platforms (Docker, Kubernetes, GCP/AWS).
• Understanding of agent design (memory management, context management, tool contracts).

Bonus Skills

• Experience with agent evaluation, prompt testing, and reducing hallucinations/latency.
• Experience building simple interfaces with Streamlit, Gradio, or browser automation.