William Amaral

Software Engineer specialized in building AI-driven systems using Large Language Models (LLMs), with experience taking solutions from research and experimentation to production.

My work includes designing context optimization strategies, implementing token usage analytics for real-time cost visibility, and developing self-healing agent systems that leverage runtime metrics to continuously improve behavior.

My work spanned from designing agentic workflows to building conversational applications and backend services. I hold a degree in Computer Information Systems from IFMG, and I'm currently pursuing a postgraduate degree in Artificial Intelligence and Machine Learning at PUC Minas.