Feb 2021 – Aug 2021 · 6 mos · Delhi, India

• Exploring Synergy with an ML based Startup

Dec 2018 – Feb 2021 · 2 yrs 2 mos · Gurgaon, Haryana

• Principal Data Scientist @PropTiger Group –
• Leading the ML team & ML initiatives for all the subsidiaries of the PT group including housing.com, makaan.com, proptiger.com, fastfox.com using state-of-the-art ML/DL techniques focusing primarily on domains such as –
• Computer Vision - classification, object detection, text detection, watermark detection, etc and their applications
• NLP – Named Entity Recognition (NER), Entity Matching, Description Generation, Language Modelling
• Predictive Modelling - Price Prediction models

Apr 2018 – Dec 2018 · 8 mos · Noida, Uttar Pradesh

• Lead projects on –
• Recommender System :-
• Recommending similar Products & Blog Articles for a consumer product company based on Factorization Machines and embedding based hybrid approaches such as LightFM
• Order Forecasting :-
• Predicting the daily online orders for a supermarket store based on history of orders data available;
• forecasting orders per slot and total daily orders for better inventory management
• Approaches such as LSTM, Statistical Time Series Analysis based approaches(ARIMA, SARIMA)
• Magic Mirror :-
• Worked on a CNN based model solution for Image Based Recommendations depending on the
• various categories of clothing worn by the user(s) standing in front of the mirror
• Worked on ResNet50, MaskRCNN, Image Classification, Object detection, Object Segmentation,
• Person Detection, Face Detection, Gender Recognition

Aug 2016 – Mar 2018 · 1 yr 7 mos · Gurgaon, Haryana

• Evaluated the problem on Generic as well as domain dataset (short text as well as generic text; structured as well as unstructured data) using supervised as well as unsupervised approaches
• Used Word2Vec, FastText feature generation for syntactic as well as semantic word relation generation
• Worked on various approaches being used for NER – CRF based models, Deep Learning based
• models( LSTM , RNN, Bi-directional LSTM-CRF)
• Explored the significance of using Regular Expressions for finding Entities
• Worked on Subject-Predicate-Object( SPO ) Models
• Worked on Ontology based models using a Local Neighbourhood Approach

Jul 2014 – Jul 2016 · 2 yrs · New Delhi Area, India

• Our services -
• 1. Corporate Training - BrainGym
• 2. Recruitment Outsourcing
• 3. Career Acceleration for students & working professionals

Feb 2013 – Aug 2014 · 1 yr 6 mos · Noida, Uttar Pradesh

• Advanced Research, Computer Vision & Machine Learning Team –
• Hand Gesture Recognition, in collaboration with IIT Kanpur
• Used ANN for localization of the target region
• Trained & Tested using SVM on various features such as – Structure Tensors, Hessian, Eigen Values/Vectors, Gabor Filter, Gaussian Derivatives
• Fourier Descriptors for shape matching
• Point Cloud Processing and error reduction for 3D video viewing using depth camera(Microsoft Kinect), in collaboration with IIT, Delhi
• Surveyed and applied Graph-cut Algorithms for noise correction and hole-filling in the raw point cloud data generated by Microsoft Kinect
• Attention Meter
• User attention modeling using frontal-face posture detection using HAAR classifier
• Eye tracking using ASEF
• Mean Shift based face tracking
• Other Research & Study
• Studied and worked on Eye Tracking, Gesture recognition for hand, Eye Detection, etc, Algorithms.
• Skeleton Finding and tracking it using Kinect X360
• Worked on Hand Tracking and Gaze Tracking Algorithms for commercialization into the 2014 D-TV

Jul 2011 – Jan 2013 · 1 yr 6 mos · Noida

• TDAA(Target Detection and Area Analysis), project on Satellite Image Processing
• Systems Development Life Cycle(SDLC), High-Level, Low-Level Design and UML
• Diagrams. Image Processing Techniques such as Filters, Hist. Specification, etc.
• Developed the GUI Architecture.
• Undertook an intern who worked on GUI, Satellite Image Import/Export,
• creating various documents such as use-case, test and review documents. GDAL
• Library used for Image Handling

Jul 2010 – Jul 2011 · 1 yr · Noida

• TRAZER – Traffic Analyzer(PRODUCT); Processes a video/live feed for detecting & classifying
• vehicles
• Worked on the following aspects –
• o The various Algorithms being used in the classification processes(Computer Vision based) for both the night and the day modes
• o HAAR classifier training and testing using OpenCV