Vikash P.

Vikash P.

Data Scientist

Hyderabad , India

Experience: 5 Years

Vikash

Hyderabad , India

Data Scientist

50400 USD / Year

  • Start Date / Notice Period end date: 2020-09-12

5 Years

Now you can Instantly Chat with Vikash!

About Me

Hi Sir/Mam,

My name is Vikash Kumar Prasad, and I have 5(Intern 1 year and 4 years as a Data Scientist), years of experience as a Data Scientist, based out of Hyderabad, India, in a couple of product-based organiza...

Products and Tech Stack worked and developed on:

  • Gainsight Sally(A conversational AI with the tech stack of  RASA-NLU+Python+Pandas+Numpy+Scipy+Jenkins+Git)

  • Cold-start-Classification(A Deep Learning based multi-classification for text, for GEP, we have filed patent also for it)

    • Python+Pandas+Numpy+Pytorch+AzureML+AzureDF+Scikit-Learn+AzureBlob

  • Newsfeeds(Automated collection of news from around the world for specific terms, finding topics and summarizing the news, Risk Modelling Using The Articles)

    • Python+Pandas+Numpy+AzureML+AzureDF+Gensim+Scikit-Learn+AzureBlob

  • Parametric Cost Modelling(A regression-based tool to ascertain cost based on different parameters and previous data history.)

    • Azure ML+Python+AzureBlob+Scikit-Learn(ML)+AzureBlob

  • Recommendation Engine(A recommendation engine used to recommend relevantly suppliers using NLP, and purchasing history)

    • Azure ML+Python+AzureBlob+Scikit-Learn(ML)+AzureBlob

  • Multiple Churn Predictive models as a Consultant Data Scientist

 

GEP Experience 

Responsibilities at GEP include understanding business problems, connecting to key stakeholders(UI/PMG), defining the flow of data collection, processing, and storage. Within a short period of time at GEP, I have developed and delivered key products that are directly proportional to the organization’s revenues. For GEP identifying new suppliers for millions of different products is of utmost importance, and hence NLP based products become of highest priority, I have tried and delivered such products in every sphere of GEP.

Gainsight Experience:

I have worked extensively on exploration and building churn predictive models at Gainsight which is primarily a CRM based organization. I have worked as a Consultant Data Scientist and build predictive models(Churn prediction) in my internship days with 5-6 customers. One of the ingredients that I learned in building predictive models was to learn how to generate effective insights that have a business impact. I developed and deployed conversational AI  Gainsight Sally, a Conversational AI. We developed a NER for this purpose using RASA-NLU and achieved an accuracy of 90% on entities. We released the chatbot back in April 2017 and it has been a success since then.

Regarding my education, I am a graduate of  IIT(ISM)-Dhanbad in M.Phil(Applied Mathematics) and M.Sc(Mathematics & Computing) from the same.

Apart from the tech stack and projects, I am concentrated on finding a good work-life balance and a company that invests in its employees. Hope I get to hear from you guys soon.

Regards

Show More

Portfolio Projects

Description

  • The objective was to gather news around clients competitors/shipping locations/suppliers in an automated form and to determine the risk around them by doing the following:

    • Summarize the news(Deep Learning-based summarization)

    • Categorize the news(Transformer models-based classification)

    • Sentiment around the news

    • Give an overall weighted score with these informations

  • Tech stack used:Python+Pandas+Numpy+AzureML+AzureDF+Gensim+Scikit-Learn+AzureBlob+Pytorch

Patent in progress

Show More Show Less

Description

The objective was to build and deploy a model that takes into account minute specifications for a large system and forecasts prices for it.

  1. Large dataset handling using pyspark
  2. Algorithms used: Bagging, Boosting

Show More Show Less

Description

  • Large Scale Text Classification For Spend Analysis:

    • Built and deployed a text classification system that would take in a line description and classify it into the various classes. The project encompasses building a whole grammar(word_embeddings+character_embeddings) for the text since most of the text is unstructured(ex: “1.5” electrical contactor safety”), we recently filed a patent.

    • Tech Stack: (Python, Azure ML, Azure Data Factory, Azure DevOps, CI/CD, Pytorch, Docker, Scikit-Learn)

    • Algorithms: (ConvNets)

Patent filed

Show More Show Less

Description

  • Implemented NLU Layer for Chatbot(Sally Gainsight) on slack. The work involved end to end deployment of the NLU(Entity and Intent recognition) Layer for chatbot on slack

    • The objective problem was to solve the NER problem for structured queries, our goal was to deliver a day-to-day interaction chatbot for CSMs/Execs/Sales so that the execs get most from basic queries for which they have the hassle to log in.

    • Tech Stack: : (Python, Jenkins, Github, RASA-NLU, Elasticsearch, Scikit-Learn)

    • Algorithms: (BiLstm+CRF, CRF)

Show More Show Less

Description

Attained a rank of 13 on LB and was called for presentation:https://www.hackerearth.com/challenges/hiring/ericsson-ml-challenge-2019/leaderboard/

Show More Show Less

Description

Analyzed and Build Model using BigQuery and Scikit Learn

Show More Show Less

Description

Meta learning

Show More Show Less