Platform Engineer based in the US.
Currently building production grade ML inference systems for real-time fraud
detection and risk scoring at Visa AI Platform. Expert in ML Infrastructure, Network, Kubernetes, Distributed Systems,
ML Lifecycle solutions and Developer Productivity
A proof-of-concept ML platform with independent model management and serving APIs.
ML model deployment requires complex ML infrastructure and specialized expertise. Built with Rust (Axum), React, and deployed on Google Kubernetes Engine (GKE), this solution provides a dead-simple API to deploy models in under a minute, powered by Rust's performance.
This project serves as an end-to-end template for Data/ML engineering using PySpark, Docker, and Kubernetes on Azure. The PySpark code processes the datasets, trains and evaluates a Linear Regression model, returns the predicted total bill, coefficients, intercept, R-squared, Mean Absolute Error, and Mean Squared Error. The Python code is dockerized, and deployed to an Azure Kubernetes Services (AKS) cluster.
A Flask CRUD web app for blogging - a place for personal musings and discussion related to tech. A platform where I also get to document my journey as a software engineer. Features: User Registration & Authentication, User Ability to Create, Read, Update, Delete posts, User Ability to update profile picture, username, email and password. Built with Python, Flask, JavaScript, SQLAlchemy, SQLite, HTML, CSS, Bootstrap. Deployed to Linode server with Nginx using Gunicorn.
Web Scraper supplemented with a game logic written in two Python scripts. First one sends a GET REQUEST to retrieve the HTML of a 10-paged website using requests. Parses, locates, scrapes 100 quotes and author info with BeautifulSoup. Writes and structures scraped data to a CSV file using the csv module.
Script two reads and appends the scraped data to a data structure called List. Initiates a guessing game that grabs a random quote from the List and asks user to guess the author. User gets 4 tries. For every failed guess, a hint is given until all attempts have been used up.
FIFA Ultimate Team is my ultimate guilty pleasure. Its transfer market is as captivating as the stock market, if not more. I built a bot that tracks player price fluctuation on FUTBIN, locates and retrieves price data as string, converts data type to float. Performs a comparison check to the price point set by the user. Takes screenshots, attaches image and price data to an email sent to user(s). The script runs on a scheduler. Time, player and email are set by the user. Utilized Python, Selenium, smtplib, MIMEMultipart, MIMEImage, MIMEText.
Read, structure, manipulate and compute 45,211 bank account data using Pandas' DataFrame & Series with marital status, education, job, account balance as key variables. Utilize Matplotlib's PyPlot to create his- tograms, bar charts, scatterplots to visualize aggregate results.
Conducted a three day project-based Python Web Scraping tutorial at Long Island University POST teaching Graduate Students (Master's) from their Data Analytics department introductory Python, web scraping, data ingestion, best coding practices, robots.txt technical syntax, ethical and legal scraping practices.