Prerak Srivastava

Prerak Srivastava

Research Scientist at SAP | P.h.D

Nancy, France

About Me

Hi! I defended my PhD in October 2023 under the supervision of Emmanuel Vincent and Antoine Deleforge at INRIA, Nancy. The main focus of this PhD was to built Machine learning systems applied to audio signals in order to estimate acoustic room parameters and sound source localization. This research further branched out to niche topics such as Acoustics, Signal Processing, Bayesien Inference and Deep Neural Network. Prior to this, i completed my Masters in Natural Language Processing from Université de Lorraine, Nancy, this masters programe was cumulated with a 6 month Research Internship at Orange Labs, Rennes, in this brief time we manage to get a acceptence for our reserach paper submitted to WASPPA 2021. So by profession i am researcher. On a daily basis, i understand and build systems to analyze audio, speech and text using machine learning techinques.

Research Interests

Audio Signal Processing, Acoustics, Natural Language Processing, Bayesian Inference, Deep Neural Networks, Machine Learning

Skills

ML/AI Libraries

Pytorch, Tensorflow, Sklearn, Scipy, ONNX Runtime

Programming Languages

Python, Matlab, C++, Bash Scripting, Latex

Tools

Git, Cython, Matplotlib

Languages

English (Fluent), French (Professional), Hindi (Native)

Experience

Data Scientist-Researcher

SAP Labs (Sophia-Antipolis), January 2024 - Present

Doctoral Research

INRIA (Nancy), September 2020 - October 2023

Improved realism in audio simulations and studied its impact on supervised learning algorithms focused on the application of Sound Source Localization and Room Parameter Estimation

Supervised By : Emmanuel Vincent and Antoine Deleforge

Research Intern

Orange Labs (Rennes), March 2020 - September 2020

Self-attention based DNN architecture to improve Sound Source Localization

Supervised By : Srdan Kitic and Pierre-Amaury Grumiaux

NLP Intern

LORIA (Nancy), October 2018 - June 2019

Anamoly detection with DNN models

Supervised By : Christophe Cerisara

Education

PhD in Computer Science (Signal Processing and ML/AI)

Université de Lorraine (Nancy)

2020-2023

MSc in Natural Language Processing

Université de Lorraine, IDMC (Nancy)

2018-2020

BE in Electronics & Communication

Government Engineering College, Ajmer (Rajasthan)

2014-2018

Publications And Patents