Lalon A.

About Lalon A.

Principal Software Technical Program Manager | ML Inference Runtime & System SoftwareI specialize in the hardware–software boundary, leading complex engineering programs that govern how ML models execute at runtime. With over 12 years of experience—including a decade at Google—I bridge the gap between high-level product objectives and low-level system execution.My approach is inference-first. I don’t just manage pipelines; I own the programs that drive P95/P99 latency optimization, determinism, and memory pressure management. I am at my best when I am the accountable DRI for the \"triage-to-fix\" loop, partnering hands-on with kernel, compiler, and runtime engineers to diagnose regressions and ensure production readiness.Areas of Impact:Inference Excellence: Scaling LLM execution, runtime scheduling, and memory budgeting for global-scale services.Systems Leadership: Navigating the complexities of first-gen accelerators (GPU/TPU/NPU) and managing hardware-software tradeoffs.Operational Scaling: Building lightweight,\"startup-speed\" delivery systems (RAID, quality gates, playbooks) for globally distributed teams (NA, EMEA, APAC).Technical Foundation: Deep roots in embedded systems and mission-critical software (Boeing, Finance), providing a first-principles understanding of reliability and performance.I am passionate about building predictable, high-velocity engineering cultures where technical rigor meets scalable execution.

Experience

1. Senior Ai Ml Tpm

   Google

   May 2014 — Present · Mountain View, CA, US

   I leveraged AI and machine learning to elevate key products, focusing on enhancing user engagement, expanding audience reach, increasing value, and retaining customers.AI Enhancements in Google Products- Google Search & Silver Search: Improved search accuracy and user experience with NLP and LLMs. Integrated AI chatbots for intuitive interactions, expanding audience engagement- Google Play Store: Personalized app recommendations using machine learning, increasing user value and retention. Employed Federated Learning for data analysis to refine recommendations- Gboard: Advanced text prediction and translation with RNNs and NLP, improving global user efficiency and widening reach. we used federated learning for on device learning- Google Area 120 (A120): Led innovative AI projects, focusing on Q Learning for new product development, targeting diverse user needs- Google YouTube (YT): Optimized content recommendations with CNNs and machine learning, enhancing user engagement and retention.Strategic AI Leadership and Agile Development- AI Chatbot Development: Developed chatbots enhancing customer service, increasing engagement and retention- Q Learning Integration: Improved decision-making and system performance across products, enhancing customer value- Agile Methodologies: Adapted quickly to technological advancements, ensuring products meet evolving user needs and market trends.Data-Driven Product Evolution- AB Testing: Systematically refined features, aligning with user preferences, increasing value and retention- Machine Learning and Algorithms: Steered product development in line with AI advancements, addressing market and user needs- Product Development: Created AI-driven, user-centric solutions, anticipating market demands, driving innovation, and expanding audience reach.

Education

• Stanford University

  Certificate, IOT product management

• Massachusetts Institute of Technology

  Applied Cyber Security