Create your own Devfolio?
mokshjain

Moksh Jain

I'm broadly interested in Machine Learning and it's applications to various domains of Computer Vision and Natural Language Processing. I'm also interested in Blockchain and Decentralised Applications.

I'm driven by my curiosity to learn about the machines that make up the modern world. I am passionate about leveraging new technologies to build innovative solutions to long-standing real-world problems. I love to build clean interfaces and robust services.

The most complex project I have worked on is Secure Aadhaar. It is a blockchain based digital identification system, to replace the existing inefficient Aadhaar infrastructure. The main reason to leverage a blockchain for this is the characteristics of transparency and security guaranteed by the blockchain. The user gets to provide selective access to his details. Each request and access is a transaction on the blockchain. This makes the process much more transparent and secure. This application was built using Hyperledger Composer for building the smart contract for the Hyperledger Fabric blockchain and the user client was built with Angular.

I worked on this project during a hackathon and I learned a lot of things. Apart from the tech stack, I was able to learn how to work effectively with a team, rapid prototyping and how to come with an idea and execute it in a limited timeframe.

I believe that hackathons are a place where passionate and talented people can come together and solve problems that have the potential to augment the daily lives of millions of people. This ability to make a difference motivates me to take part in hackathons.

Projects

#BUIDL FOREST

Imagine Pokemon Go, but a whole lot greener, without the fluffy monsters. #BUIDL FOREST motivates you, with the help of cryptotrees, to go out and help the environment off-chain.Vue.js, Leaflet, OpenZeppelin

VideoDex

Content based video indexing engine which analyzes the visual content in the video and builds a searchable index. Place banner advertisements optmially and search through videosDjango, PyTorch, spaCy, SQLite

Skills

Python
Java
JavaScript
PyTorch
Hyperledger Composer
Keras

Experience

  • Eureka King Inc San Francisco - Machine Learning Intern
    May 2018 - July 2019

    Implemented deep recurrent models for financial time series forecasting improving performance over state of the art models by 5% and worked on reinforcement learning algorithms for optimizing trading portfolios in highly volatile financial markets, yielding improvements in daily profits of over 2%.

  • Leuphana Universität Lüneburg - Research Intern
    May 2019 - July 2019

    Developed a novel algorithm to extract interpretable player and team rewards from given expert demonstrations of professional football players using based on inverse reinforcement learning.

  • Microsoft - Research Intern
    August 2019 - Present

    Part of the EdgeML team working on developing efficient machine learning algorithms suitable for inference on resource constrained devices.

  • for.ai - Researcher

    Working on a Tensorflow framework for asynchronous training of reversible component based neural networks on TPUs and curiousity driven approaches for exploration in reinforcement learning