Building AI Agents with AWS Strands SDK - Nova Month Challenge
A walkthrough of completing all 5 AWS Strands Agents challenges using Ollama, Amazon Bedrock Nova Pro, MCP tools, and persistent memory.
I completed all 5 challenges in the AWS User Group Madurai Builders Skill Sprint (Nova Month) using the Strands Agents SDK.
Challenge 1 - First Agent: Built a basic agent using Ollama (llama3.2:3b) locally. Added a pirate mode bonus with a custom system prompt.
Challenge 2 - Tools: Connected the agent to a calculator, weather API (wttr.in), age calculator, and a km-to-miles converter. Demonstrated two tools firing in one query.
Challenge 3 - Memory: Used mem0_memory with FAISS vector store to persist user preferences across sessions. Stored name, city, and favourite food.
Challenge 4 - Full Agent: Combined all tools and memory into one interactive chat agent running on Amazon Nova Pro via Bedrock.
Challenge 5 - Innovate (MCP): Built a Dev Productivity Agent using the filesystem MCP server. The agent reads local code files, lists directories, and performs code reviews - all powered by Nova Pro.
Key takeaway: Strands makes it surprisingly simple to wire up tools, memory, and MCP servers to a Bedrock model in just a few lines of Python.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article