Varun Jewalikar

London, UK

Summary

I’m a software engineer with 6 years of experience of delivering large scale distributed systems and machine learning systems (audio, text and image based). I believe in learning the right tool for the job.

Skills

Interests

Backend engineering, machine learning, natural language processing, data visualization, audio analysis, web scraping, rapid prototyping.

Tools

Java, Kotlin, AWS, Javascript, Python, HTML, CSS, C, AWS, Scikit-learn, Linux, Git, Bash.

Languages

English, Spanish, Italian, Hindi and Marathi.

Experience

Software Engineer, Amazon

2017 - Present
London, UK

I design, implement and maintain high traffic services (1 million requests/sec) for Prime Video. These services are deployed on a mix of AWS and Amazon internal infrastructure.

Designs done by me have saved $2M YoY.

Java, Kotlin, Javascript, Python, Spring, DynamoDb, Cloudformation, Ec2, AWS IoT

Consultant and tutor

Oct - Dec 2016
London, UK

I consulted/taught Python, machine learning, natural language processing and software engineering to professionals and students.

Software Engineer, Stylezz

Mar - Oct 2016
London, UK

I designed the backend for this shopping app. I prototyped deep learning based clothes categorization and image-image search system to improve discovery.

Python, MongoDb, Flask, Scikit-learn, Caffe, Ec2, Elastic beanstalk

Software Engineer, Musixmatch

2014 - 2016
Bologna, Italy

I did audio/lyrics/data analysis projects, collaborated with researchers, hired/supervised interns and made features for products.

I made data driven articles which been featured in The Guardian, Billboard.com, etc. Other research projects: mood prediction, improving lyrics sync, detection of new releases from Youtube and instrumental song detection.

I travelled to 20 hackdays, demonstrated our APIs, helped hackers and built things.

Python, Javascript, MongoDb, Scikit-learn, Flask, NodeJS, Webpack, NLTK, Ec2

R&D trainee, Yamaha Music

Sep - Dec 2013
Japan

Vocaloid is a singing voice synthesizer. I used ML to improve phonetic segmentation for the creation of new singing voices for Vocaloid.

Python, Scikit-learn, Git

Intern, Mixxx.org

Apr - Jul 2013
Barcelona, Spain

I was a Google Summer of Code intern for Mixxx which is an open source DJ-ing software written in C++. I added key detection, tempo independent pitch shifting and harmonic mixing for DJs.

C++, Qt, Git

Intern, Beagleboard.org

Apr - Jul 2010
Delhi, India

I was a Google Summer of Code intern for Beagleboard which is a low-power single-board computer by Texas Instruments. I wrote a high level interface for Pulse width modulation via the output pins.

C, Git

Education

M.S. in Sound and Music Computing, Universitat Pompeu Fabra

2011 - 2013
Barcelona, Spain

Selected courses: Audio & music processing, Music perception & cognition & Machine Learning.

Thesis: Improving automatic phonetic segmentation for creating singing voice synthesizer corpora.

B.E. in Computer Engineering, Delhi College of Engineering

2006 - 2010
Delhi, India

Thesis: A novel parallel clustering algorithm implementation with CUDA.