Henrico, VA • 802-777-1255 • aravind.pk@hotmail.com
View My profile @ GitHub | LinkedIn | StackOverflow | Pluralsight
Mar 2026
Card Transact Advisor combines classical ML (K-Means, Linear Regression, weighted scoring) with a 29-
feature engineering pipeline and Anthropic's Claude Haiku LLM to deliver real-time card
recommendations, health scoring, anomaly detection and plain-English financial insights — all
personalized to each user's actual spending behavior.
This was a final capstone project as part of my Bootcamp program from Virginia Tech.
Backend: Python FAST API | Frontend: React with Vite | Database: PostgreSQL | LLM: Anthropic Claude Haiku | Feature Engineering: Scikit-learn
Mar 2024 – Present
This will read the USCIS visa bulletin page and provide the key dates by visa type and country. Renders on a nice Web app to display the dates with options to select. Having next set of road maps to this.
Python with WebScraping and Flask with elementary Html for web
Oct 2025 – Present
This is a personal interest, .net Aspire based project started as learning but with a real world application scope as a goal to accomplish. The app would download the 990 filings made publicly available by IRS and renders an intuitive web UI displaying the every year filings Data.
.NET Core, C#, Aspire, Redis with (mongo/Postgres Yet to be integrated)