- Spearheaded the design and development of a contract analysis application, leveraging Streamlit, Azure App Service, and Azure OpenAI. This application revolutionized contract review processes by providing insightful recommendations for modifications.- Led a critical evaluation of a deployed computer vision application in a manufacturing environment, utilizing Azure Machine Learning and Azure IoT Edge. The assessment revealed scalability issues, prompting a comprehensive overhaul of the system for improved performance and efficiency.- Collaborated with a leading electronic component and services provider to build a cutting-edge LLM agent POC, designed to extract valuable insights from SEC documents. Additionally, provided a robust architecture blueprint to facilitate scaling to an enterprise application.- Coordinated with an international development team to automate the process of producing a statement of work. We built a FastAPI backend, utilizing Azure Cognitive Search and Azure OpenAI. This initiative streamlined the creation of comprehensive statements of work, specifically tailored to pre-sales input.

- Developed a POC for a metadata-driven framework on the Databricks platform aimed at streamlining the training and deployment of machine learning models.

- Developed a POC for integrating Google Vertex AI and Kubeflow Pipelines with flagship application.- Utilized Cloud Build and ArgoCD to deploy Streamlit dashboard applications to Google Kubernetes Engine.- Redesigned resource hierarchy and security policies on Google Cloud Platform using the principle of least privilege.

- Ported an existing API to Google Cloud Platform running containers in Cloud Run and ML models in AI Prediction.- Prototyped an essay auto-scoring system using AI Platform custom training pipelines.- Collected requirements and researched platform solutions for automating data scientist’s ML workflows.

- Benchmarked and deployed TensorRT Triton Server to host deep learning models and perform accelerated inference.- Built a scheduling chatbot using Node.js and natural language processing to automate customer appointment creation.- Installed, configured, and managed monitoring tools including Prometheus, cAdvisor, Node Exporter, and Grafana.