Daivya Shah

At Newmark, I worked on several web applications used by brokers and research teams to manage real estate data and deals. One of my main projects was a Commercial Real Estate stacking plan that visualized tenant and vacancy data floor by floor. I built it with React and TypeScript, adding filters, interactive layouts, and color-coded lease expirations to make portfolio insights easier to interpret.

I also developed a time-based activity feed that merged notes, tours, proposals, and deal events into one view, using React Query to handle live data updates and caching. Later, I built a Kanban-style deal pipeline with drag-and-drop deal tracking, inline editing, and dynamic columns to help brokers organize their workflows. Toward the end of my internship, I designed a responsive GIS landing page with Tailwind CSS that showcased site selection services through modular components and integrated form submissions. Working across these projects helped me strengthen my understanding of scalable component design, state management, and how small UI changes can meaningfully improve daily workflows.

At Ubiq, I worked as a Product Management Intern where I focused on improving user experience and overall product performance. I led user research by conducting 60+ customer interviews and turned those insights into 5 key UX improvements that directly enhanced usability and engagement.

I also executed full-cycle product testing, identifying 50+ bugs and collaborating with engineers to resolve 4 critical issues affecting core functionality. Additionally, I managed an agile backlog of 350+ features, UI enhancements, and bug fixes in Jira, prioritizing development based on user needs and business impact.

At Everise, I worked on automating data workflows and improving how performance metrics were tracked. I built ETL pipelines using Python, SQL, and Apache Airflow to combine data from different client systems into a central warehouse. Once the data was integrated, I created Power BI dashboards with DAX to visualize key metrics like Bill-to-Pay percentage and service-level performance. Debugging inconsistencies across different data sources was difficult, but it helped me learn how to manage reliability in production data systems. I also developed and deployed a Go-based forecasting API that used ARIMA and LightGBM to predict client-level Bill-to-Pay rates with around 90% accuracy. This experience gave me a clearer understanding of data engineering processes and how automation improves business reporting.

At eMeasurematics, I worked on predictive maintenance software for industrial vehicles in steel facilities. I helped develop a Flask-based system that used SQLAlchemy, MySQL, and ensemble models such as Random Forest and XGBoost to predict maintenance needs from sensor data. Cleaning inconsistent readings and tuning models for stable performance was one of the harder parts of the project. I also created a Tableau dashboard to monitor machine performance and configured radar sensor firmware used in crane positioning systems. The experience taught me how to apply machine learning to real operational problems and how reliable data collection directly affects the accuracy of predictive models.