Shravya S Madhusudan

Assistant Professor,

Department of CSE, PES University

about me

Aspiring Computer Science Researcher, specialised in Artificial Intelligence, Machine Learning and Natural Language Processing. With a research focus on advancing regional language technologies, Computer Vision and Human-Computer Interaction through interdisciplinary and impactful innovation.

Education

PES University
Master of Technology, Computer Science & Engineering (Artificial Intelligence)

Visvesvaraya Technological University
Bachelor of Engineering, Computer Science & Engineering

Skills 

Programming Languages 

Python, C, JavaScript 

Augmented & virtual reality

Unity 3D, Three.js, 3D environment 

Machine Learning

Linear Regression, Decision Trees, Random
Forest Regression, Support Vector Machines. 

Deep learning

Convolution Neural Networks, Probabilistic
Neural Networks, Recurrent Neural Networks,
GANs.  

Natural Language
Processing

Natural Language
Processing

Text Classification and Sentiment Analysis

experience

experience

PES University, Bengaluru, India

Assistant Professor, Department of CSE
(January 2023- Present)

Teaching undergraduate courses in Machine Learning, Python, and C Programming, while supervising research projects integrating ML, Computer Vision, and Natural Language Processing, and conducting hands-on labs to bridge theory with practical application.

Member of the research centers CAVE (Centre for Augmented and Virtual Environments) and ISFCR (Centre for Information Security, Forensics and Cyber Resilience), contributing to research in AI, AR/VR, and Human-Computer Interaction, and mentoring students on mixed reality and AI-based projects.

Organized departmental hackathons and workshops to promote innovation and collaborative problem-solving among students.

NetApp, Bengaluru, India 

Member Technical Staff (Intern)
(October 2022- July 2023)

Developed a data analysis project to predict the Recovery Point Objective (RPO) based on various data points in different customer environments. The project involved evaluating the effectiveness of the configured RPO in the client environment using reduced RPO/schedules.

RESEARCH AND PUBLICATIONS 

UK Design Patent (Filed: Nov 2025): Vision-Based Road Pattern Recognition Device Using Image Processing. 

Indian Patent (Filed: Oct 2025): System and Method for Optimising Corporate Decision-Making Using AI-Driven Analytics. 

UK Design Patent (Filed: Aug 2025): Smart Farming System Using IoT and Cloud Integration. 

Journal Submission (Applied: Sep 2025): Augmented Reality–Assisted Robotic Surgery Using Real-Time Image Processing and 3D Reconstruction. 

Publication: Rice Grain Quality Determination Using Probabilistic Neural Networks, in Sustainable Communication Networks and Applications, Springer, DOI: 10.1007/978-981-15-8677-4_21. Link 

Publication (Accepted): Virtual Reality in Hotel Marketing: Enhancing Pre-Booking Experience, International Journal of Integrated Research and Practice (IJIRP), ISSN: 3107-5037, DOI: 10.25215/31075037 Link

Conference Paper (Applied, Oct 2025): Auto Image Generation Using Kannada Text Descriptions, submitted to an international conference. 

Conference Paper (Applied, Oct 2025): Speech-Driven Question Answering System for the Kannada Language, submitted to an international conference 

Download CV

Research projects 

Hybrid Rope Physics Engine: For 3D simulation

Developed a comprehensive rope physics simulation framework by implementing a custom Verlet integration model alongside Ammo.js and Cannon.js physics engines. Created real-time, interactive 3D visualizations that support smooth movement, gravity-based dynamics, and 360° free rotation of the rope. Designed the system to compare performance, stability, collision accuracy, and computational efficiency across the three simulation methods. Built fully functional demos demonstrating particle constraints, damping, tension control, and rope behavior consistency under different forces. This project showcases expertise in numerical physics modeling, engine-level simulation, and visually accurate real-time rendering using Three.js.

Controlling Adulteration in Public Food Grain Distribution

Developed an ML-based classification system that uses the characteristics and physical parameters of grains to detect adulteration in public food distribution channels. Implemented a Probabilistic Neural Network (PNN) to accurately differentiate pure and contaminated grain samples based on texture, color, and pattern features. The system supports data-driven quality assurance, reducing manual inspection errors and improving fairness in public distribution systems. Designed the project to increase transparency and reliability in food safety processes. Demonstrated strong potential for scalable deployment in government-supported food monitoring pipelines.

Speech-Driven Question Answering System

Built an end-to end spoken question answering system integrating ASR (Automatic Speech Recognition) with QA (Question Answering) specifically for the Kannada language. Focused on improving accessibility for low-resource language communities by enabling voice-based information retrieval. Enhanced the ASR pipeline to manage dialectal variations and noisy speech conditions common in real-world settings. The QA module retrieves accurate answers from structured knowledge sources using NLP and semantic search. This work contributes to regional language AI, improving inclusivity and digital access for native speakers.

Auto Image Generation Using Kannada Text Descriptions

Developed a multilingual text-to-image generation pipeline using Diffusion Models, GANs, and NLP feature encoders. Trained the system on Kannada textual prompts paired with corresponding images to enable meaningful visual synthesis. Achieved 83% accuracy, demonstrating the feasibility of generative AI for lowresource regional languages. The model supports both creative image generation and educational applications in multimodal Kannada datasets. This work highlights innovation in combining generative AI + NLP for Indic languages. Generative Adversarial Networks (GANs) were used to refine image realism and improve visual consistency across generated outputs.

Toxic Comment Detection Using NLP

Designed an NLP-based system to identify and classify toxic or abusive language in online text. Used supervised machine learning algorithms and linguistic feature extraction to detect harassment, hate speech, and offensive content. Pre-processed text using tokenization, lemmatization, and contextual embeddings to improve classification performance. The system strengthens digital safety by supporting automated moderation in social media and chat platforms. Focused on improving recall for harmful content to ensure maximum protection for online communities.

Multi-Threaded Chat Application

Built a fully functional multi-threaded chat server enabling real-time, low-latency communication among multiple clients. Implemented mutex-based synchronization to ensure thread safety and avoid race conditions during message broadcasting. Optimized socket-based communication to support scalability and stable performance under heavy client load. The system demonstrates strong foundations in computer networks, concurrency handling, and distributed communication. Designed with extensibility in mind, allowing future upgrades like encryption, authentication, and file sharing.

Virtual Reality in Hotel Marketing

Conducted research on using VR-based immersive environments to enhance customer engagement before hotel booking. Developed interactive 3D hotel walkthroughs to allow users to experience rooms, amenities, and ambience virtually. Showed that VR significantly improves decision confidence and user satisfaction, especially for premium hospitality brands. The study integrates concepts from HCI, UX design, and experiential marketing, providing actionable insights for the industry. Published the findings to demonstrate how VR can transform digital marketing strategies in the hospitality sector

Vision-Based Road Pattern Recognition Device

To add a video to your site, click the “Insert” button and navigate to the “Media” section. Then, drag and drop a video component onto the Canvas.

Smart Farming System Using IoT and Cloud Integration

Designed an IoT-enabled smart farming system that collects environmental data such as soil moisture, humidity, and temperature. Integrated cloud-based analytics to monitor crop health and provide automated alerts for irrigation and fertilization. Improves agricultural efficiency by enabling data-driven decision-making for farmers, especially in resource-limited areas. The system uses sensors, microcontrollers, and dashboard visualization for real-time monitoring. Supports scalable agricultural automation and contributes to sustainable precision farming practices

Adventurous Game Prototype Development Using Virtual Reality

Created a VR-based adventure game featuring immersive 3D environments, realistic physics, and motion-controlled interactions. Implemented user navigation, object interaction, and gameplay mechanics using Unity and C programming. Designed the prototype to focus on player engagement through spatial audio, dynamic lighting, and intuitive motion gestures. Explores game development within VR for entertainment, simulation training, and experiential learning. Showcases strong capability in blending graphics, HCI, and real-time interaction within virtual worlds.

Hybrid Rope Physics Engine: For 3D simulation

Developed a comprehensive rope physics simulation framework by implementing a custom Verlet integration model alongside Ammo.js and Cannon.js physics engines. Created real-time, interactive 3D visualizations that support smooth movement, gravity-based dynamics, and 360° free rotation of the rope. Designed the system to compare performance, stability, collision accuracy, and computational efficiency across the three simulation methods. Built fully functional demos demonstrating particle constraints, damping, tension control, and rope behavior consistency under different forces. This project showcases expertise in numerical physics modeling, engine-level simulation, and visually accurate real-time rendering using Three.js.

Controlling Adulteration in Public Food Grain Distribution

Developed an ML-based classification system that uses the characteristics and physical parameters of grains to detect adulteration in public food distribution channels. Implemented a Probabilistic Neural Network (PNN) to accurately differentiate pure and contaminated grain samples based on texture, color, and pattern features. The system supports data-driven quality assurance, reducing manual inspection errors and improving fairness in public distribution systems. Designed the project to increase transparency and reliability in food safety processes. Demonstrated strong potential for scalable deployment in government-supported food monitoring pipelines.

Speech-Driven Question Answering System

Built an end-to end spoken question answering system integrating ASR (Automatic Speech Recognition) with QA (Question Answering) specifically for the Kannada language. Focused on improving accessibility for low-resource language communities by enabling voice-based information retrieval. Enhanced the ASR pipeline to manage dialectal variations and noisy speech conditions common in real-world settings. The QA module retrieves accurate answers from structured knowledge sources using NLP and semantic search. This work contributes to regional language AI, improving inclusivity and digital access for native speakers.

Auto Image Generation Using Kannada Text Descriptions

Developed a multilingual text-to-image generation pipeline using Diffusion Models, GANs, and NLP feature encoders. Trained the system on Kannada textual prompts paired with corresponding images to enable meaningful visual synthesis. Achieved 83% accuracy, demonstrating the feasibility of generative AI for lowresource regional languages. The model supports both creative image generation and educational applications in multimodal Kannada datasets. This work highlights innovation in combining generative AI + NLP for Indic languages. Generative Adversarial Networks (GANs) were used to refine image realism and improve visual consistency across generated outputs.

Toxic Comment Detection Using NLP

Designed an NLP-based system to identify and classify toxic or abusive language in online text. Used supervised machine learning algorithms and linguistic feature extraction to detect harassment, hate speech, and offensive content. Pre-processed text using tokenization, lemmatization, and contextual embeddings to improve classification performance. The system strengthens digital safety by supporting automated moderation in social media and chat platforms. Focused on improving recall for harmful content to ensure maximum protection for online communities.

Multi-Threaded Chat Application

Built a fully functional multi-threaded chat server enabling real-time, low-latency communication among multiple clients. Implemented mutex-based synchronization to ensure thread safety and avoid race conditions during message broadcasting. Optimized socket-based communication to support scalability and stable performance under heavy client load. The system demonstrates strong foundations in computer networks, concurrency handling, and distributed communication. Designed with extensibility in mind, allowing future upgrades like encryption, authentication, and file sharing.

Virtual Reality in Hotel Marketing

Conducted research on using VR-based immersive environments to enhance customer engagement before hotel booking. Developed interactive 3D hotel walkthroughs to allow users to experience rooms, amenities, and ambience virtually. Showed that VR significantly improves decision confidence and user satisfaction, especially for premium hospitality brands. The study integrates concepts from HCI, UX design, and experiential marketing, providing actionable insights for the industry. Published the findings to demonstrate how VR can transform digital marketing strategies in the hospitality sector

Vision-Based Road Pattern Recognition Device

To add a video to your site, click the “Insert” button and navigate to the “Media” section. Then, drag and drop a video component onto the Canvas.

Smart Farming System Using IoT and Cloud Integration

Designed an IoT-enabled smart farming system that collects environmental data such as soil moisture, humidity, and temperature. Integrated cloud-based analytics to monitor crop health and provide automated alerts for irrigation and fertilization. Improves agricultural efficiency by enabling data-driven decision-making for farmers, especially in resource-limited areas. The system uses sensors, microcontrollers, and dashboard visualization for real-time monitoring. Supports scalable agricultural automation and contributes to sustainable precision farming practices

Adventurous Game Prototype Development Using Virtual Reality

Created a VR-based adventure game featuring immersive 3D environments, realistic physics, and motion-controlled interactions. Implemented user navigation, object interaction, and gameplay mechanics using Unity and C programming. Designed the prototype to focus on player engagement through spatial audio, dynamic lighting, and intuitive motion gestures. Explores game development within VR for entertainment, simulation training, and experiential learning. Showcases strong capability in blending graphics, HCI, and real-time interaction within virtual worlds.

Get in touch 

location

Bengaluru, India

Contact details

shravyamadhusudans@gmail.com 

+91-8867808606

Linkedin

location

Bengaluru, India

Contact details

shravyamadhusudans@gmail.com 

+91-8867808606

Linkedin

location

Bengaluru, India

Contact details

shravyamadhusudans@gmail.com 

+91-8867808606

Linkedin