San Jose, California, United States
🌐 About Hi, I’m Johnson Lo, an ECE Master’s student at Carnegie Mellon University passionate about building scalable AI and cloud solutions. I thrive at the intersection of cloud architecture, machine learning, and real-time systems. Easy to work with, I bring both technical depth and a collaborative mindset. ☁️ AI + Cloud Expertise From deploying Kubernetes-based systems at TSMC to developing GenAI evaluation frameworks at Scale AI, I’ve built secure, scalable platforms across diverse environments. At Google Nest, I engineered an edge-cloud pipeline on GCP for real-time IoT monitoring, boosting data accuracy by 19%. I’m also AWS Solutions Architect – Associate and Google Cloud Certified Associate Cloud Engineer, showcasing proven cloud skills. 🤖 Applied AI Across Industries I’ve delivered AI-powered financial advisory tools on AWS for HKS Partners, and designed an LLM-enhanced public service robot for Honda Research using NLP, cloud APIs, and edge computing. These projects translated into measurable business impact—like reducing deployment time by 55% and cutting latency by 30%. 🚀 Entrepreneurial & Systems Thinking Beyond internships, I co-founded an AI-driven automotive safety startup that won a national competition and sold its patent for $10,000. With experience spanning distributed systems, DevOps pipelines, and multi-agent LLM research at CMU, I’m driven by systems thinking and innovation. 🤝 Let’s Connect! I’m always excited to collaborate on projects at the frontier of cloud infrastructure, AI systems, and product innovation. Let’s explore how we can build something impactful together.
Fine-tuned LLMs with curated test suites and structured rubrics, boosting accuracy in code-generation, logic reasoning, and complex problem-solving by optimizing training data and evaluation signals. Engineered scalable evaluation pipelines in Python for Olympiad-style math, logic, and code reasoning tasks, detecting edge-case failures and strengthening model reliability with automated testing. Authored production-grade prompts and golden responses to automate QC workflows and scale structured output validation.
Building a decision-making agent driven by large language models in RL environments, integrating observation prompts, action inference, and environment feedback to study how LLMs reason and act within multi-agent systems.
Developed a secure, web-based Kubernetes file browser with user-centric design and robust access controls, using Java, Python (SpringBoot, FastAPI), Angular, React, and Azure, improving system access efficiency by 40% and ensuring compliance. Automated configuration management by developing a tool to parse Kubernetes YAML files and manage dependencies in MongoDB, optimizing system updates and reducing downtime by 35%. Managed the full software development lifecycle, implementing Agile methodologies and CI/CD pipelines using Kubernetes and Azure, enhancing deployment efficiency by 30% and system scalability.
Conducted 5G network data analysis, applying Python (Pandas, NumPy, Scikit-learn) to extract insights from RF performance metrics, optimizing signal transmission and reducing power consumption by 33%. Developed predictive models for 5G network optimization using MATLAB and Python, improving frequency allocation efficiency and accelerating product deployment by 20%. Developed and validated RF component simulation models using C++ and Ansys HFSS, enhancing 5G hardware design and accelerating product deployment by 20%.