At TLScontact, I developed and deployed time-series forecasting models for visa application volumes and sales, reaching 96% and 94% accuracy. I also delivered a forecast for Canadian visa applications that was used during client negotiations.
I built data pipelines and automated workflows with Snowflake and Microsoft Power Automate, and created Matillion jobs to transform data. I designed four Power BI and Metabase dashboards and monitored more than 20.
At TLScontact, I applied clustering to segment services into four groups, supporting targeted campaigns that increased conversion rates by 2–10%. I also led machine learning training sessions and collaborated with external partners.
At STB Bank, I built and evaluated ATM cash-demand forecasting models that achieved 90–92% accuracy. I also implemented web scrapers to enrich the data, and created a banking chatbot that reduced customer support issues by nearly 30%.
