• Designed and optimized scalable, fault-tolerant distributed systems, improving latency, availability, and resource efficiency in Linux-based environments.

• Developed high-performance system software in C/C++ and Python, optimizing memory management, process scheduling, and I/O operations.

• Engineered multi-threaded and parallel computing solutions, reducing compute overhead and improving real-time processing.

• Debugged low-level system issues, profiling CPU utilization, memory leaks, and kernel interactions using GDB, pdb, and dtrace.

• Implemented load balancing and distributed caching strategies (Redis, Memcached, ccache, distcc), reducing latency and optimizing compute efficiency.

• Architected containerized and virtualized environments with Docker, Kubernetes, and KVM, ensuring resource isolation and efficient orchestration.

• Integrated monitoring and observability frameworks (Prometheus, Grafana, ELK Stack) for real-time system health tracking and anomaly detection.

• Developed custom networking solutions, optimizing TCP/IP stack performance, packet processing, and secure communication.

• Developed a data structures and algorithms visualization app for University of the Pacific, aiding 100+ students in understanding complex algorithms.

• Mentored 350+ students in DBMS, data structures, operating systems, and computer networks, assisting with error-debugging and coding in C/C++, Python, and JAVA.

• Collaborated with firmware and software teams to implement smart-meter and smart grid system using C++, Arduino IDE, and NGINX.

• Recommended cloud services based on use case needs, saving costs and improving efficiency.

• Implemented test automation using Selenium and performed unit-testing on large server data.

• Evaluated hardware capabilities and suggested efficient ways to connect smart meters, reducing costs by 15%.

• Developed an Automatic carparking system using ROS, Object Detection, and SLAM for Hewlett Packard Enterprise.

• Implemented a convolutional neural network-based vacant parking detector with OpenCV and YOLO modules.

• Translated Python code to C++ to enhance the embedded system's runtime by 39%.