Hi there! I am Ayushi!

Welcome to my data science playground! I’m a curious, creative, and coffee-fueled data enthusiast who loves uncovering patterns, building smart solutions, and turning messy data into meaningful insights. Whether it’s segmenting customers, forecasting logistics, or detecting fast radio bursts from deep space, I enjoy using machine learning and statistics to solve real-world problems that matter.

This portfolio is a collection of projects that reflect both my technical skills and my love for data storytelling. From predictive models and dashboards to end-to-end pipelines and ML experiments — every project here has taught me something new and pushed me to think smarter.

When I’m not deep in a dataset, you’ll probably find me hosting tech meetups, geeking out over new Python tools, or daydreaming about building a model that predicts when I’ll finally get 8 hours of sleep.

Thanks for stopping by — I hope you find something here that sparks your curiosity too!

Photo by Cash Macanaya on Unsplash

Retrieval-Augmented Employee Analytics Using LangChain and Gemini

In this project, I built a GenAI-powered assistant to explore synthetic HR data through interactive dashboards and natural language queries.

Photo by Jakub Zerdzicki on Unsplash

Predicting Loan Outcomes: A Deep Dive into Risk Modeling

A machine learning project to predict loan defaults using real-world LendingClub data and risk modeling techniques.

Photo by SumUp on Unsplash

Data-Driven Customer Segmentation for an Online Retailer

In this project, I performed customer segmentation on retail transaction data to identify key customer groups for targeted marketing.

An artist’s illustration of artificial intelligence (AI). This illustration depicts language models which generate text. It was created by Wes Cockx as part of the Visualising AI project launched by Google DeepMind.

Token and Entity Explorer with BERT and PyTorch

In this project, I trained a BERT model using PyTorch to tokenize input sentences and accurately identify named entities, exploring how transformer-based models perform sequence labeling tasks.

CNN based real-time classification pipeline

This project is on the development of an end-to-end state-of-the-art ML based FRB detection pipeline for the UTMOST radio telescope in python using Keras and TensorFlow.

Galaxy Zoo CNN Classification

In this project, Galaxy zoo images were used to train and test a network to classify elliptical and spiral galaxies.

Kickstarter project success prediction

In this project, featues of a Kickstarter project such as the category, country, currency, the pleded amount (in USD and in the given currency), the deadline for the completion of the project is used to predict the state of the project (whether it is successful or not).

Whale Detection using audio clips

A collection of 30,000 audio files was sourced from the Marinexplore and Cornell University Whale Detection Challenge to classify clips containing whale sounds from those without any.

Photo by Liam Truong on Unsplash

Certifications

Get in touch with me!

  • Address

    Melbourne
    (Send me a message if you'd like to get some coffee and chat about some of the latest papers in ML!)

  • Email

    ayushimandlik09@gmail.com

  • Phone

    (+61) 481015577

  • Social