Pranav Sharma

Experience

Graduate Researcher - Fraunhofer IIS

Jan 2022 - Sep 2024 | Erlangen, Germany

Developed and deployed the CINGO spatial audio system in Unity 3D, enhancing the user experience in virtual environments by creating modular software for real-time spatial audio rendering.
Conducted a comparative analysis of loudspeaker and headphone recordings using sine sweep, analyzing Room Impulse Responses (RIRs) and correlating acoustic measures with perceptual audio quality.
Recorded loudspeaker/headphones using sweeps for multiple source-receiver configurations to improve and enhance the RIR dataset.
Designed Acoustic Visualization tools using Qt/C++, enabling room acoustics visualization and analysis of large datasets.
Enhanced Binaural Room Impulse Responses (BRIRs) and assessed innovative spatial audio plugins, optimizing data-driven audio systems for improved localization, timbre, and envelopment, which ties into automated data extraction and performance analysis.
Played a pivotal role in the Mobile Audio Rendering team, enhancing overall user experience in spatial audio using signal processing.
Executed MUSHRA tests to assess spatial audio perceptual effects with HRTF across diverse populations, achieving a 30% improvement in audio quality.
Designed and implemented an analysis framework for comparing various spatial binaural renderers, enabling comprehensive comparison and optimization of spatial audio tools.
Created a RIR similarity matching algorithm using unsupervised machine learning, automating the comparison of Frequency Domain Network (FDN)-based RIRs with real-world RIRs, improving system efficiency by streamlining data processing.

Associate Engineer - Tata Consultancy Services

Sep 2018 - Jun 2020 | Pune, India

Designed and deployed a three-tier web application using Docker, Kubernetes, Jenkins, and AWS EC2, improving deployment efficiency by 40%.
Deployed a complete environment, including VPC, EC2 instances, RDS, and S3 buckets, with version-controlled templates.
Implemented a CI/CD pipeline using AWS CodePipeline and CodeBuild, automating the deployment of a serverless application with AWS Lambda and S3, reducing deployment time by 30%.
Maintained and optimized the Model Data Repository for offshore banking operations with the Bank of Montreal, ensuring high data integrity and performance.
Designed data models and implemented complex SQL queries for data extraction, reporting, and analytics, reducing query execution time by 20%.
Developed a RESTful API for managing database interactions, integrating DynamoDB for highly available data processing.
Standardized infrastructure deployment using AWS CDK and CloudFormation templates, reducing manual configuration errors by 80%.
Deployed a production-grade Kubernetes application on AWS EKS, optimizing resource allocation and scaling by 50%. Created Helm charts to streamline deployments and manage versioned updates, achieving seamless rollouts with zero downtime. Automated cluster provisioning and scaling through CI/CD pipelines using GitHub Actions and Terraform, reducing deployment cycles by 60%. Ensured high availability with 99.95% uptime and improved operational efficiency by cutting manual intervention by 45%.

Projects

AI-Based Decay Parameter Estimation

Master's Thesis

Developed a novel DCNN model for estimating decay parameters directly from reverberant speech signals, achieving RMS errors of 2.9 to 6 dB for EDC loss and 0.08 to 0.21 seconds for reverberation time (RT) loss.

GitHub Link

Spotify Trendz - Real-Time Music Trend Analysis & DevOps Automation

Real-Time Music Trend Analysis

This project automates real-time music trend analysis using Kubernetes and Terraform. By leveraging EKS for orchestration and Docker Swarm for container management, it reduced deployment time by 40%. Scalability and fault tolerance were enhanced through container orchestration, with Prometheus and Grafana integrated for monitoring and observability.

GitHub Link

Spring Boot CI/CD Pipeline

Automated Deployment Pipeline

A robust CI/CD pipeline for automating the deployment of a Spring Boot application using Jenkins, Docker, and Kubernetes. Integrated SonarQube for static code analysis, deployed in a Kubernetes cluster with GitOps managed by ArgoCD, and containerized using Docker.

GitHub Link

Three-Tier Web Application

Digital Resume Application

A three-tier web application implemented as a Dockerized microservice designed to act as a digital resume. The application uses a frontend built with HTML, CSS, and JavaScript and a backend Django application. Deployed in Kubernetes for scalability and reliability.

GitHub Link

AI/ML Pipeline

Developed and deployed a fully automated machine learning pipeline, utilizing Elastic Beanstalk for serverless application deployment. Integrated AWS CodePipeline with GitHub for CI/CD, automating the deployment process and enabling seamless model training, evaluation, and deployment.

GitHub Link

Real-Time Contour Detection

Developed a two-tier web application for real-time contour detection and sensor control using Raspberry Pi and OpenCV. Deployed the system using AWS Elastic Beanstalk and Docker.

GitHub Link

Video Processing

Created Python programs for channel segregation, quantization, DCT, kernel filters, object detection, and pickling for video processing tasks.

GitHub Link

Context-Based Image Retrieval

Implemented a system to retrieve images from a large database based on their visual content using machine learning techniques.

GitHub Link

VR Game Development

Developed an arm-swinging locomotion game using HTC Vive, enhancing VR user experience by reducing dizziness by 60% while playing games.

GitHub Link

Loudspeakers to Go

Created a mobile virtual loudspeaker setup on headphones using the Binaural Room Impulse Response database and Intel RealSense depth camera with 3 DoF, enhancing the spatial audio experience on portable devices.

GitHub Link

Audio Processing

Developed Python programs for down/up sampling, filtering, quantization, low delay filter-banks, pitch/speed enhancement, DCT, MDCT, and QMF for advanced audio processing tasks.

GitHub Link

Summary

Skills

Certifications

Experience

Graduate Researcher - Fraunhofer IIS

Associate Engineer - Tata Consultancy Services

Projects

AI-Based Decay Parameter Estimation

Spotify Trendz - Real-Time Music Trend Analysis & DevOps Automation

Spring Boot CI/CD Pipeline

Three-Tier Web Application

AI/ML Pipeline

Real-Time Contour Detection

Video Processing

Context-Based Image Retrieval

VR Game Development

Loudspeakers to Go

Audio Processing

Education

Master's in Media Technology

Bachelor's in Electronics and Communication