May 2017 – May 2018 · 1 yr · Raleigh, North Carolina

• Developed Image classifiers and detectors using caffe based deep neural networks for multiple categories.
• Improved precision for 2 classifiers and 1 detector by almost 10-20% by identifying issues in data and classes and by improving the training parameters.
• Improved the automation of training and testing pipeline to significantly reduce the manual intervention.
• Developed an Anomaly Detection system for an alert system product to detect anomalies in time series data using HTM (Numenta) and generate meaningful alerts to be sent to customers using Kafka.
• Image Detection. Tensorflow and caffe based detectors.

May 2017 – Sep 2017 · 4 mos · Raleigh-Durham, North Carolina Area

• Managed the core code repository and the Django server hosting the application for open source TEMOA Energy Model. Restructured code and improved runtime for UI visualizations by more than 70%.
• Added ability to generate new visualizations on the front end, automated the process of setting up the application on a new server and improved front end interface. (http://model.temoacloud.com)
• Worked with Prof. Joseph DeCarolis

Mar 2016 – Jun 2016 · 3 mos

• Worked on creating C# bindings for Tizen APIs for Sensor Framework and Audio Manager

Jun 2014 – Feb 2016 · 1 yr 8 mos

• Open Source Contributions in LLVM Compiler Infrastructure. (July 2014 - Nov 2014)
• Contributions in Tizen Open Source (Sensor Framework) (Dec 2014 - Dec 2015)
• Worked on framework optimizations, restructuring and addition of various features. Improved template design for various classes, updated queue implementation for event handling of sensors.
• All contributions at: https://review.tizen.org/gerrit/#/q/owner:iankurgarg

May 2013 – Jul 2013 · 2 mos · Bengaluru Area, India

• I worked on using Latent Dirichlet Allocation methods for categorization of Social and News Feeds of users to make it easier for a user to traverse through the various Facebook Posts, Tweets, News Articles, RSS Feeds etc.

May 2012 – Jul 2012 · 2 mos · Hyderabad Area, India

• As an intern here, I worked on optimization of compiler for Cavium's Neuron Search Processor. Worked on reducing the runtime of compiler with large datasets. Reduced the runtime of the compiler code by about 10 %.