Building LLM Agents with Tool Use
summary
Demonstrates building an LLM-backed agent that connects Cohere Command R+ to external tools via LangChain and a web-search API (Tavily) plus a Python interpreter. The agent plans, performs four web searches to find 2024 full‑time headcounts for Microsoft, Apple and NVIDIA, generates matplotlib code, executes it, and returns a plotted result with source snippets. Shows multi‑hop tool use, runnable notebook, and successful end‑to‑end code execution.
AI-written summary · about
chapters (6)
- 0:00Introduction to tool use
- 1:04Concept of LLM agents
- 2:16Data analyst agent example
- 3:16Setting up agent tools
- 4:13Executing the agent workflow
- 6:52Conclusion and possibilities
related talks
41mBuilding LLM Agents – Challenges and Learnings
35mBuilding LLM-based Agents
01h 43mBuilding LLM Agents (Hackathon)
25mLLM Powered Agents with Kubernetes
30mBuilding Robust, Secure LLM and Agentic AI Workflows
23mLLM Agents in Production: Why Go is the Missing Piece
01h 10mTeach an LLM to Navigate the Web
01h 04mBuilding LLM powered applications in Ruby