AI Agent Developer – Tool-Calling and Function Protocol Design
Architected and implemented modular tool systems for AI agents which included designing function calling protocols for LLMs. Developed structured inputs/outputs for tool-calling agents, enabling AI to perform actions like file editing, code execution, and web browsing. Maintained accuracy and utility of toolkit function definitions for AI agent benchmarking and improvement.• Created datasets for tool-calling demonstration and function selection training • Documented and iterated over function schemas for LLMs • Developed test cases and evaluation sets for tool-calling accuracy • Supported multi-agent tool interoperability and data consistency