Ananya Muralidhar
Ananya Muralidhar
COMPUTER SCIENCE ENGINEER
RESEARCH INTERESTS
I am passionate about harnessing data-driven solutions for societal betterment with a particular interest in applying AI in healthcare. I've explored this interest by working on projects that have exposed me to the various challenges faced in this field such as data imbalance, noise, explaianability, biases, generalization error and volume.
I want to further my research in these domains to build user-centric intelligent artificial systems that are robust, interpretable, and fair, designed not only to diagnose problems but also to accurately recommend and substantiate a course of action, especially essential in high stake sectors such as healthcare.
I want to explore the following areas:
1. Explainable AI (XAI) in Healthcare: Delving into transparent AI models to decipher their decision-making mechanisms, ensuring both clinicians and patients have clarity and confidence in AI-driven healthcare recommendations.
2. Decision-making Enhancement using LLMs: Harness large language models to bolster system decision-making, introducing cognitive adaptability to handle complex scenarios.
3. Adaptive Information Systems: Dynamically integrate and interpret vast data sources, ensure interoperability, security, and a smooth user experience, and provide real-time analytical insights for early and personalized interventions.
PROJECTS
Automated Depression Detector From Speech
Created a sophisticated automated tool designed to detect depression from prosodic features in speech for early detection and treatment of depression. The model is trained on over 3000 minutes of interaction from USC's DAIC-WOZ Database.
Secured Fog-Based System for Smart Healthcare Application
Implemented an encrypted three-tier (user, fog, cloud) architecture with a Random Forest Classifier to assess patient symptoms and recommend timely medical actions on an intuitive web interface.
Analysis of RNN Capsule Model for Emotion Classification
Evaluation of the performance of ML models (Decision Tree, Random Forest, Naive Bayes, SVM) alongside RNN-capsule for multiclass imbalanced data from the Twitter Emotions Dataset.
Bixby, Capsules, Marketplace | Out-Domain Utterance Identification
Developed an advanced classifier for Samsung's Bixby to identify out-of-domain utterances, enabling the voice assistant to determine and respond appropriately to inputs not belonging to any supported capsules.
Client-Server Application for Missing Persons Management
A structured client-server application designed to address challenges in recording and managing missing persons, offering users a platform to efficiently interact with the option to view, add, or track information related to missing individuals.
Library Management System
A Java-based system streamlining library operations by digitally cataloging items and offering users a user-friendly web interface for efficient borrowing and returning, with robust database integration and RESTful API capabilities.
Jun 2022 - Present
Mar - Jun 2022
Sept 2021 - Apr 2022
Jun - Aug 2020
Software Engineer
Merkle Inc.
-
Specialise in Data Engineering, Natural Language Processing and Continuous Development and Integration of applications to improve data retrieval efficiency and expedite client responsiveness.
-
Work as a cross-functional team member to optimize workflows, processes, and system architectures, improving operational efficiencies while embracing Agile methodologies for multiple development teams.
-
Create comprehensive technical documentation and an industry pitch deck, used as a standardizing framework for project execution and client presentations across Merkle and our parent company, Dentsu.
Data Analytics Consultant Intern
Merkle Inc.
-
Performed market analysis by comparing revenue and gross margin trends for leading US-based retail distributors, pinpointing key drivers and formulating targeted pricing strategies for increased revenue increase across retailers.
-
Automated the web scraping process for insights into user behavior, optimizing marketing and enhancing user experience by developing scripts for web interaction, cookie extraction, and dynamic JSON configuration.
Samsung R&D Institute India
Machine Learning Intern
-
Worked on using in domain utterances as the dataset to build an AI/ML model which can detect out domain utterances with better accuracy for integration into Bixby.
-
Implemented multiple multiclass and binary classification models across various embedding layers and compared their accuracies, noting the impact of each type of model and contribution of datasets in building the optimum model.
-
Awarded the 'Certificate of Excellence' given to the top performers of the program
Ecotex Environmental Technologies
Database and Systems Intern
-
Designed and Implemented a comprehensive database of the company’s projects to improve document control and project management.
-
Conducted in-depth machinery failure analysis isolating the most common-causes of failure and presented insights via Tableau, reducing downtimes by 15%.
PROJECT
IEEE
Institute of Electrical and Electronics Engineers
Driven by a passion for leveraging technology for social good and a dedication to promoting inclusivity in STEM, my journey with IEEE provided a substantial platform to manifest these ideals.
​
Serving in dual leadership roles—unprecedented in the history of IEEE RIT—I've had the incredible opportunity to lead, manage, and innovate various initiatives which have allowed me to make impactful changes in these avenues on the lives of those around me.
​