Shivam Pasricha

Roles :

• Use machine learning and statistical modelling techniques to uncover insights in data sets.
• Worked on horizontal scaling of python using Ray framework.
• Use visualisation tools like Kibana and Redash to give data insights.
  Demonstrate applications built, to the prospective clients.
• Develop applications for the customers, learning new technologies and enabling those technologies in FORMCEPT's product.

Projects :

1. SMS Data extraction and classification:

• Modules used : Keras (LSTM), Regular Expressions.
• Data labelling using Regex to differentiate between Transactional, Promotional, Shopping, Travel etc SMSes.
• Extract meaningful data from the SMS using Regex example (Product name, expected delivery data, actual deliver data in case of shopping message).
• Implemented LSTM model to classify the new SMS with 95% accuracy.

2. Time series component:

• Modules used: Statsmodels, Ray, Flask.
• Plotting time series data to extract decomposed components of data points (Seasonal, Trend Residuals).
• Finding unusual data points using statistics (Running Mean and Standard deviation).
• Time series forecasting using ARIMA model provided number of steps to forecast.

3. ID Card Classification:

• Modules used : Keras (CNN).
• Applied CNN to classify different ID card like Pan card and Driving License.

4. ETL (Extract Transform Load)

• Tools Used: Spark
• Get data from different data sources
• Large data pre-processing and Loading data to Elasticsearch using Spark.

5. EVSE - CMS (Electric vehicle central management system):

• Tools : Kibana , X-pack.
• Implemented central management system for Electric Vehicle using OCPP Protocol.
• Used Kibana and X-pack graph for visualisation of data and generate insights from the data.

Projects :

1. Peepin

Description :This application helps to peep inside Bars. It shows 10 sec video clips of bars. Also It give all the information of bars like music, menu, events and crowd.

Tools : Appsflyer for deep-linking for customised application url ,Helpshift SDK for customer support services, FCM for push notification Paytm SDK for payment.

2. Productive One

Description: This application helps to monitor the usage of different applications in android device.And can take screenshots of android device in the background.

Tools: Accessibility Service API for fetching the package information of foreground application, FCM for push notification, Media Projection API for fetching the screenshots of device.

3. Jeenees :

Description: This app let's user to create events and add 3 wishes for that event using the flipkart products.User’s friends acts as jeenees and helps to fulfil the wishes by contributing money for the wishes.
Tools : Branch.io for deep linking, Facebook SDK for social login, CcAvenue for payment gateway