Qira: Enterprise-Grade Agentic AI

I’m an AI product leader focused on operationalizing AI, turning model capability into reliable, user-centered outcomes across real products, devices, and workflows. My strength is bridging technical ML understanding with product execution and governance, so AI works in the real world of edge cases, latency, risk, and multi-team delivery. Over the last 1–3 years at Google, my work centered on AI trust, safety, and risk governance, helping validate and evaluate AI-enabled software. This required hands-on technical judgment across machine learning behavior, policy and privacy requirements, and failure-mode analysis, including defining evaluation criteria, reviewing risk controls, and ensuring systems could be assessed consistently and responsibly at scale. I helped ensure AI behavior was not only high-performing, but also accountable, testable, and defensible. More recently, I joined Lenovo to lead agent ecosystem work, including Qira—an ambient AI assistant experience designed to help users move from intent to outcomes across day-to-day productivity and cross-device contexts. My focus is making agentic AI trustworthy and usable: tight interaction loops, transparent system behavior, and repeatable patterns that avoid one-off “demo magic.” I work across engineering, UX, and cross-functional stakeholders to drive clarity, alignment, and execution readiness. A key part of this work is pushing beyond traditional “chat” into agentic, context-aware systems: assistants that can understand what the user is doing, synthesize information into decisions, and safely execute multi-step tasks when appropriate. Practically, that means designing reliable tool-orchestration patterns, building evaluable workflows (not just prompts), and creating the UX that makes AI feel predictable—clear “thinking/working” states, explicit confirmation points, and concise outcome summaries. This is bleeding-edge work because it combines model reasoning, product constraints, and real-world integration while maintaining guardrails and user trust. I’m also a named inventor on a Dynamic Context-Aware Teleprompter for Smart Glasses—an AI-driven interface that synchronizes speaker notes to active content and adapts visibility based on speaking and context signals, reducing cognitive load and improving delivery quality. This nomination recognizes my individual contribution to building AI that is not only impressive, but deployable, trustworthy, and designed to create real user value.