4Bell Technology

Job Description

Key Responsibilities

• Architect multi-agent systems using frameworks such as LangGraph, AutoGen, CrewAI, or custom orchestration layers, with a focus on reliability and observability.
• Design and implement tool-use pipelines — including function calling, MCP integrations, browser agents, and code interpreters — that enable models to interact with real-world systems.
• Build memory systems (short-term, episodic, and semantic) and context management strategies to support long-horizon task completion.
• Define and run evaluation frameworks for agentic performance: task completion rates, error propagation, cost-efficiency, and latency benchmarks.
• Collaborate with product, research, and data teams to translate business goals into concrete agent specifications and acceptance criteria.
• Establish guardrails, safety checks, and human-in-the-loop escalation paths to ensure agents behave predictably in edge cases.
• Monitor deployed agents in production using tracing tools (LangSmith, Weights & Biases, custom observability stacks) and iterate based on failure analysis.
• Stay current with the rapidly evolving agentic AI landscape and advocate for adoption of relevant advances internally.

Required Skills

• Python (advanced) LLM APIs Agent frameworks RAG & vector stores
• Tool / function calling Prompt engineering REST & async APIs Docker / cloud
• Evaluation design Git & CI/CD