I am the Lead Software Engineer at UBC Agrobot, a design team focused on designing autonomous agriculture robots and their support systems. The team repositories can be seen here. Some of my notable contributions as team lead can be seen below:

• Optimized model deployment, improving inference speed by 5x (from 15 to 75 FPS) through Nvidia-based quantization and GPU acceleration, enabling real-time video processing
• Increased Machine Learning team productivity by 59% by developing internal Python tools to streamline ETL workflows and reduce data processing time
• Streamlined release times by 63% by building and maintaining CI/CD pipelines using GitHub Actions to eliminate manual build and deployment operations
• Accelerated sprint delivery by 46% by automating unit, integration, and performance testing workflows with Python and Bash scripts, reducing human intervention

• Created a GoLang CLI tool to enable seamless SSH container access across local and remote environments, including user authentication and real-time status updates through gRPC services

As a cloud engineer intern at UBC, I performed a migration of backend infrastructure to the cloud, improving resource allocation for scalability and adding optimizations across the API, database and service queue layers. Here is a summary of my biggest accomplishments:

• Saved $8000+ in operational costs and computing resources by migrating backend infrastructure to the cloud using Terraform and AWS resources (VPC, EC2, RDS, Lambda, S3)
• Reduced manual querying time by 10+ hours per week by building FastAPI-based endpoints to integrate a JavaScript frontend with ElasticSearch, enabling quicker and more accurate data retrieval
• Achieved a 74% reduction in latency (from 200ms to 52ms) for upstream services by designing a real-time log streaming solution using Apache Kafka and object-oriented programming
• Improved data retrieval times by 62% by integrating Redis in-memory caching with a PostgreSQL database for efficient user data storage

• Decreased response times by 55% by developing a GoLang application to improve notification visibility on Slack using GitHub APIs and webhook payloads

I formerly worked on the Internal Tools subteam for UBC Agrobot, where I improved the robot communication architecture and developed various tools to improve developer productivity, my primary achievements include:

• Minimized system-wide latency from 2.1s to 0.1s by developing and deploying C++ components using ROS2 to improve network communication within the weed extermination robot
• Decreased environment setup times by 37% using custom multi-platform Docker containers to automate local and staging setup procedures, eliminating dependency issues for 20+ developers
• Improved application performance monitoring and reduced issue resolution time by implementing comprehensive monitoring and alerting systems using Prometheus and Grafana

• Reduced project deployment times by over 71% using Helm IaC templates to automate microservice architecture configuration on Kubernetes clusters across 5+ multiple availability zones

• Created a collection of Python scripts for automating common tasks, including web scraping, data analysis, and file manipulation, saving 10 hours of manual work per week.