Upon joining as the Tech Lab Administrator at the premier lab in Eastern and Southern Africa equipped with GPUs, I assumed the critical responsibility of enhancing our software capabilities. My role involved enabling staff to test and validate various applications within a real-world environment, thereby driving business value.A notable project during my tenure was the deployment of Rancher in collaboration with the infrastructure team. This open-source platform, poised to save the… Show more Upon joining as the Tech Lab Administrator at the premier lab in Eastern and Southern Africa equipped with GPUs, I assumed the critical responsibility of enhancing our software capabilities. My role involved enabling staff to test and validate various applications within a real-world environment, thereby driving business value.A notable project during my tenure was the deployment of Rancher in collaboration with the infrastructure team. This open-source platform, poised to save the company billions in costs upon full implementation, was tested rigorously in our environment. My contributions included provisioning Virtual Machines, installing the Rancher version, and setting up a Kubernetes Cluster to manage containerized workloads efficiently.A key deliverable in my role was ensuring robust monitoring capabilities. I spearheaded the monitoring of host health, CPU, memory, and storage metrics. Together with the Service Reliability Team, we explored and utilized advanced tools such as Dynatrace and the ELK Stack (Elasticsearch, Logstash, and Kibana). Leveraging my data analytics expertise, I developed and deployed a comprehensive dashboard, capturing and visualizing metrics collected twice daily. This involved utilizing a Linux environment to access the virtual machine on the server, dockerizing my Streamlit application using Podman, and deploying it seamlessly into the server. Show less

- Performed Exploratory Data Analysis using Python(Pandas, Matplotlib) on the Demand and Ticketing with a view to ultimately creating an efficient system of detecting degradation, that is, predictive analytics - Retrained a Supervised Machine learning model utilizing Supervised Machine Learning Models on Pyspark, to improve performance by 10% from the existing Demand Classifier Model. - Performed SQL queries to pull data from PostgreSQL Database.