• Developed and implemented a request management system using llama.cpp for LLM inference in a multi-user environment, optimizing resource allocation and request prioritization.• Integrated llama.cpp to enable high-performance inference for large language model deployment across two distributed clusters (16 AMD GPUs), utilizing ROCm and HIPBLAS acceleration.• Contributed to the llama.cpp open-source project by identifying a performance improvement that prioritized RPC servers in the device list, reducing network transfer of logits and significantly enhancing inference performance.

• Developed custom analytical tools in Python and VBA to automate provincial energy data processing andvisualizations for 1000+ delivery points, enhancing the efficiency of energy distribution monitoring.• Developed and implemented advanced statistical models in Python for energy demand forecasting.• Managed data validation and editing efforts, systematically identifying and resolving data discrepancies.

• Developed a real-time monitoring system using Arduino and multiple sensors to evaluate solar wall technology.• Designed a Java application for data manipulation/ real time data visualization, data collection

• Converted an AC-DC computer power source into a DC input power source for integration into a DC microgrid• Developed and tested hardware prototypes using breadboards, oscilloscopes, multimeters, and other testing tools