Default profile banner
RN

Roshan Nayak

@user.2567816

Research Assistant | ML & Signal Processing Engineer

Bangalore, India

https://github.com

Indian Institute of ScienceB.M.S College of Engineering

Roshan Nayak is an Electronics and Communication Engineering graduate from BMS College of Engineering with a strong academic record and deep expertise in machine learning, deep learning, and signal processing. He currently works as a Research Assistant at the Indian Institute of Science, Bangalore, focusing on model compression techniques for Automatic Speech Recognition systems and air writing detection. Roshan has previously worked at Synopsys and Samsung Prism, and has published research on NLP and meme detection at ICON-2021 and SemEval-2022.

Experience

Research Assistant

Indian Institute of Science

Full-time•Jan 2023 - Present•Bangalore, India

Working on projects related to signal processing and model compression under Prof. Dr. Prathosh AP. Preprocessing Automatic Speech Recognition (ASR) datasets, fine-tuning models, and conducting inferences. Researching model compression techniques including knowledge distillation and quantization to reduce size and latency of ASR models. Also working on Air Writing detection using deep learning of sensor signals.

Research and Development Engineer

Synopsys Inc.

Full-time•Jul 2022 - Dec 2022•Bangalore, India

Developed and tested front-end features for the Formality software using C++. Resolved customer-specific bugs. Took responsibility of the handoff consumption process integrating latest dependencies into Formality tool and conducting performance analysis.

Technical Intern

Synopsys Inc.

Internship•Feb 2022 - Jun 2022•Bangalore, India

Developed VHDL test cases to conduct stress testing for the Formality software. Reported software limitations in handling complex VHDL constructs.

Project Intern

Samsung Prism

Full-time•Jun 2021 - Dec 2021•Bangalore, India

Worked on generating OCR data for eight low-resource Indian languages. Collaborated in training UNet architecture for image segmentation. Used depth estimation and image processing to place text on images realistically. Assessed dataset quality using open source OCR models.

Data Science Intern

Career Robin

Part-time•Apr 2021 - Dec 2021•Mumbai, India

Built data pipeline from scratch using Selenium and Beautiful Soup for real-time web scraping. Processed scraped data and stored in AWS S3. Developed an NLP model to extract required skills from job descriptions and hosted it on AWS EC2.

Education

B.M.S College of Engineering

Bachelor of Engineering

Electronics and Communication Engineering

Aug 2018 - Jul 2022•Grade: 9.13/10

Relevant coursework: Big Data Analytics, Machine Learning, Image Processing, Probability and Statistics, Data Structures and Applications, Java Programming, Computer Communication Networks.

Licenses & Certifications

Python Programming (3 courses)

University of Michigan

Deep Learning Specialization (4 courses)

deeplearning.ai

Software Architecture and Clean Code Design

Learn IT University on Udemy

Skills

Python
C/C++
SQL
Java
Kotlin
MATLAB
MySQL
Git
AWS
Linux
OpenCV
Scikit-Learn
PyTorch
Keras
TensorFlow
Django
Android Studio
Deep Learning
Machine Learning
NLP
Signal Processing
Model Compression
Knowledge Distillation
Quantization
Selenium
Beautiful Soup
Computer Vision