Abhinav Jain

I am a tinkerer who loves to play around with things, break and build systems. Good engineering and tech always amazes me.

I primarily am a backend developer and code in Python, Java, C++. Have had experience working with robotic systems and hardware, so ROS, Arduino, Raspberry Pi, are not new to me.

Not sure if the most complex, but the most interesting project that I worked on was during my stint (internship) at Capillary Technologies, where I was responsible for developing the backend for a router based identification devices. That involved writing Shell (Bash) and Lua scripts and setting up a pipeline (with cron jobs) for dumping network packets to a SQL database and then sending it across a network to an AWS S3 instance. There was a lot of work involved with building firmware and flashing it on these devices. This system was to be deployed in malls and public places. Also collaborated on Linux kernel development and building a user facing registration portal using Django.

I have also interned at Laboratory of Robotics and Automation at HEIG-VD (School of Business and Engineering, Vaud) in Yverdon-les-Bains, Switzerland where I worked on building an interface between ROS and Windows environments to encourage communication and collaboration between different frameworks. Also built 3D simulations of domestic helper robot on Gazebo.

Last year I was at American Express Big Data Labs, where I was part of the fraud detection team, and was working on fraud classification. Task was to explore boosting ensemble methods in the online setting, so that a model could be trained on the fly and results obtained simultaneously.

Currently, I am working as a Software Development Engineer at Flipkart, in the User Acquistion and Growth division, building products for brands, mining ratings and reviews data to generate actionable insights.

I believe I carry with me a diverse background in Software Development / Machine Learning / Robotics, and have that passion for building products that could have a positive impact on society, and that's the reason for applying to the InOut Hackathon.

Some other projects that I have worked on:

• Recommending User Topics for Questions on Quora: Automatically tag questions with relevant topics using Word Mover's Distance to get similar words using pre-trained embeddings

• Emotion Classification using Neural Networks: An application to recognize emotions from a person’s facial expressions using Haar features and Gradient Descent model fitting. Extracted features include Gabor-filtered images,and auto-generated features from pre-trained networks. Trained an artificial neural network on the Cohn-Kanade database of facial expressions achieving a high accuracy of 84.4%

• Swarm Robotics - Decentralized Terrain Exploration: Researched on path planning, motion control, and SLAM for swarm of autonomous robots.