All work

Cue 01, 2015 to 2018

Before BMAT

Before BMAT I ran infrastructure for enterprise websites in Bangalore, then moved to London for a master's in Sound and Music Computing and a research placement on sending data over sound. Through all of it, I played lead guitar in a band.

Role
DevOps engineer, then researcher
Stack
Puppet, Ansible, Docker, Elasticsearch, AWS, Azure, MATLAB, Python, C++

Infrastructure DevOps, SapientNitro, 2015 to 2017

I started in Bangalore on SapientNitro's infrastructure team, after training in Java, JavaScript and HTML, looking after the servers behind large enterprise websites.

  • Built configuration management with Puppet, and automated infrastructure with Ansible and Docker.
  • Built a log monitoring platform on Elasticsearch, Logstash and Kibana, with reporting and alerting.
  • Worked across AWS and Azure, and automated SSL certificates so they renewed themselves before expiring.
  • Set up and ran web and application servers (Apache, IBM HTTP Server, Tomcat, JBoss), including load balancing and the content management platforms on top.

MSc Sound and Music Computing, Queen Mary University of London, 2017 to 2018

I'd always wanted to combine music and engineering, so I moved to London for the MSc in Sound and Music Computing, and finished with a Distinction.

The course covered digital signal processing, real-time DSP, digital audio effects, music analysis and synthesis, music and speech modelling, music perception and cognition, and machine learning. Projects included an interactive surface projected with Kinect depth sensing, FlexDrummer, and an analysis of how the singing voice has changed over the years.

Research Technician, Chirp, London, 2018

For my master's project I joined Chirp, a company that sends data between devices as sound. The question was how rooms and other acoustic environments affect those signals, and what to do about it.

I applied my research to Chirp's signal processing stack, measured performance and prototyped new techniques. The work became a paper at the 27th European Signal Processing Conference (EUSIPCO 2019): Estimating & Mitigating the Impact of Acoustic Environments on Machine-to-Machine Signalling. It proposed a new way to estimate the acoustic parameters that cause reverberation, then used them to clean up the signal before decoding, with a large drop in error rates for both audible and ultrasonic signals.

Before that

  • BE in Electronics and Communications Engineering, PES Institute of Technology, Bangalore, 2011 to 2015. My final projects included automatic chord estimation from audio, using chroma features and a hidden Markov model.
  • Lead guitarist in 1Fret Away, Bangalore, 2010 to 2017.