Payoj Jain

Payoj Jain

Bengaluru, Karnataka, India
8K followers 500+ connections

About

Product Building | Consumer Data Analysis | Machine Learning | Data Science | Model…

Experience

  • Teachmint Graphic

    Teachmint

    Bengaluru, Karnataka, India

  • -

    Gurgaon, India

  • -

    Boulder, CO

  • -

    Boulder, CO

  • -

    New Delhi, Delhi, India

  • -

    Noida Area, India

  • -

    Noida Area, India

Education

  • Indian Institute of Technology, Delhi Graphic

    Indian Institute of Technology, Delhi

    -

    -

    Activities and Societies: Dance, Badminton, Basketball, Entrepreneurship, Web Development

  • -

    Courses: Neural Networks and Deep Learning, Recommender systems, Design and Analysis of Algorithms, Machine Learning, Natural Language Processing, Big Data Architecture, Object Oriented Analysis and Design, Validation and Uncertainty Quantification of Computational Models
    Independent Studies: Prof. James Martin : CLEAREarthNLP, Prof. Rebecca Morrison : Hydrogen and Methane Combustion Model Analysis and Performance

Publications

  • Automated Trait Extraction using ClearEarth, a Natural Language Processing System for Text Mining in Natural Sciences

    Biodiversity Information Science and Standards

    ClearEarth is a collaborative project that brings together computational linguistics and domain scientists to port Natural Language Processing (NLP) modules trained on the same types of linguistic annotation to the fields of geology, cryology, and ecology. The goal for ClearEarth in the ecology domain is the extraction of ecologically-relevant terms, including eco-phenotypic traits from text and the assignment of those traits to taxa. ClearEarth uses the NLP results to generate domain-specific…

    ClearEarth is a collaborative project that brings together computational linguistics and domain scientists to port Natural Language Processing (NLP) modules trained on the same types of linguistic annotation to the fields of geology, cryology, and ecology. The goal for ClearEarth in the ecology domain is the extraction of ecologically-relevant terms, including eco-phenotypic traits from text and the assignment of those traits to taxa. ClearEarth uses the NLP results to generate domain-specific ontologies and other semantic resources.

    See publication
  • Chhattisgarh’s food ATMs: Portable benefits minus biometrics

    India Together

    Portability has become a buzzword in recent debates on social security benefits. The idea is to create a system wherein beneficiaries of a programme are able to avail their entitlements (subsidised food, or NREGA wages or old age pensions), irrespective of their place of residence. Since a large number of social security beneficiaries are poor migrants, portability can lead to significant convenience.

    See publication

Courses

  • Analysis and Design of Algorithms

    CSL 356

  • Big Data Architecture

    -

  • Data structures

    CSL201

  • Design and Analysis of Algorithm

    -

  • Discrete mathematics

    CSL105

  • Introduction to Linguistics

    -

  • Investment Planning

    MEL323

  • Machine Learning

    -

  • Machine Learning

    -

  • Natural Language Processing

    -

  • Natural Language Processing

    -

  • Probability and Stochastic Processes

    -

  • Programming Languages

    CSL 302

  • Public Finance and Public Economics

    -

  • Topics in Cognitive Science

    -

  • Validation and Uncertainty Quantification of Computational Model

    CSCI7000

Projects

  • RNN Model Evaluation & Development for Entity Recognition

    - Present

    - Investigated various machine learning and deep learning approaches to entity recognition in earth science domain
    - Developing various metrics for named entities in three different domains to evaluate the performance of the model
    - Developing RNN model using python libraries like Pytorch and Tensorflow to improve current model performance

    See project
  • Part of Speech Tagging

    Hidden Markov Model and Viterbi Algorithm
    - Implemented statistical approach to part-of-speech tagging using over 15,000 POS-tagged sentences
    from the Berkeley Restaurant corpus as training data
    - Improved accuracy from 93.94% to 96.29% by implementing Viterbi algorithm with bigram-
    model based approach, various smoothing techniques and handled unknown words problem

    See project
  • Investing in Markets, Workshop Conducted by J.P. MORGAN

    • Among the 30 students shortlisted from 3 prestigious colleges to attend the workshop about finance basics and things which go behind the investing decisions in the markets
    • Designed and came up with a strategy in a team of 5 for allocating money among 10 underlying assets

  • Twitter Opinion Mining and Topic Modelling for Brands and Products

    -

    - NLP and ML techniques to automatically generate an aspect-based sentiment summary from real-time tweets on Twitter
    - Users can see opinions and sentiments of people for a particular product of a particular electronics brand on heats maps based on certain demographics like location, gender and age-group
    - Worked on AWS EC2, Kafka, Spark, NLTK, TwitterNLP, Python, Java

    See project
  • Healthcare Platform

    -

    - Developed a healthcare platform where people can locate doctors, pharmacists, pathologists, consultants and book appointments.
    - The portal also manages medical records/history of each patient so that it is easier for doctors to make their diagnosis and pharmacists can suggest medicines to patients.
    - Gained hands-on experience Java, Agile Development, Design Patterns, Refactoring

    See project
  • Image Depth Estimation

    -

    - Developed and compared various CNN based architectures from fully connected models to purely convolutional and also used transfer learning with pre-trained models to estimate depth map from monocular images.
    - Created the “Depth effect” (blurring the background appropriately to create DSLR like effect) using depth map output to extrinsically evaluate our depth estimates and used it to identify and differentiate between the foreground and background of an image.
    - Evaluated models on…

    - Developed and compared various CNN based architectures from fully connected models to purely convolutional and also used transfer learning with pre-trained models to estimate depth map from monocular images.
    - Created the “Depth effect” (blurring the background appropriately to create DSLR like effect) using depth map output to extrinsically evaluate our depth estimates and used it to identify and differentiate between the foreground and background of an image.
    - Evaluated models on indoor scenes and object-focused images. Started model evaluation with a fully connected network as baseline (RMSE: 1.4) and moved to fully convoluted network to achieve an RMSE of 1.32

    See project
  • Undergraduate Thesis: Digitization of Running Accounts in local stores; Kirana Khata

    -

    • Developed an Android Application to overcome the problem of change faced at local stores and shops
    • Transactions are made via Bluetooth; customers enter details of purchases; shopkeepers verify details
    • Optimised the application by reducing number of steps to one and making payment process faster

    See project
  • Predicting stock price movement; Sentiment Analysis on microblogging sites

    -

    • Constructed a robust model to forecast stocks direction by aggregating tweets posted on StockTwits.com
    • Studied sentiments of tweets for over 10 stocks; applied machine learning techniques for predictive analysis
    • Achieved an accuracy of over 75% in predicting stocks by operating on data for over a month period

  • Analysis of Public Distribution System (PDS) in Chhattisgarh

    -

    • Scrutinized performance of COREPDS after surveying 41 Fair Price Shops (FPS) in Raipur District
    • Studied working of Supply Chain Management of food grains in each Fair Price Shop by government
    • Identified bottlenecks in system and areas of improvement on supplies, sales and stocking operations
    Achievement: Co-authored report on COREPDS which got published in online journal India Together

    Other creators

Test Scores

  • All India Engineering Entrance Exam (AIEEE 2011)

    Score: 305/360

    All India Rank 285

  • Joint Entrance Exam (IIT-JEE 2011)

    Score: 378/480

    All India Rank 116

View Payoj’s full profile

  • See who you know in common
  • Get introduced
  • Contact Payoj directly
Join to view full profile

Other similar profiles

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Others named Payoj Jain in India

Add new skills with these courses