Gustavo Valente

Software Engineer with experience in production LLM systems — design, deployment, observability, and cost/latency optimization at scale.Over the past two years I helped grow an AI startup (Nero.AI) from its early days into a company serving 2M+ users across 20+ enterprise clients, with R$1M+ in AI services delivered. Selected work: - Designed and deployed high-scale LLM systems, focused on observability, reliability, and reducing inference cost and latency. - Telus Digital (AI Engineer, consulting): built a text-to-SQL + voice-to-voice system letting field agronomists safely query enterprise data, plus an internal tool that automated much of internal data-compliance workflow. - Research: combined eye-tracking with lightweight LLMs to improve assistive typing for people with motor disabilities — awarded best project in my college's 2024–25 research cycle (funded by Brazil's national research council).Background: full-ride scholarship at Insper; gold medalist, Brazilian Physics Olympiad. Currently completing a 2-month CS research program at UIUC.Core skills: LLM application development, RAG, agentic systems, text-to-SQL, voice AI, observability, cost & latency optimization, Python, TypeScript/Next.js, AWS (EKS, ECR), GCP, Kubernetes.