Mohit Garg

Experience

• 

  Mindstone Maven LLP

  Jun 2019 - Sept 2019

  Machine Learning Intern
• 

  Unyscape

  Oct 2019 - Sept 2020

  Data Analyst

  - Gained solid understanding of data analysis, data preprocessing, data cleaning and several machine learning algorithms such as Logistic Regression, Ensemble models (bagging and boosting), SVM, clustering and pipeline development for machine learning workflow.- Developed as Automated Anomaly detection tool for detecting anomalies behavior in web data. Tool was based on keywords clustering; several algorithms such as BERT, Affinity Propagation, K-mean, hierarchical clustering were tested.- Worked with an E-commerce client in analyzing, cleaning web data (Organic, Adwords, Facebook Pixel) and developed several dashboards using SQL queries and automated using GCP big query and Google Data Studio.- Automating manually generated report using python and scheduling over AWS to minimize manual workload Show less

• 

  Algoscale

  Sept 2020 - Sept 2021

  Data Scientist

  Project C- Worked with an ad-tech client and developed a system for segment website visitors and recommending them best ad running over website on the basis of their engagement data- Conceptualized and developed a Probabilistic model based on Markov model for recommending offers ads running over websites to optimize clicks/conversion.- Extensively used Amazon Web Services such as - ec2, s3, lambda, sagemaker for deployment, maintenance and A-B testing.- Recommendation model is evaluated using A-B testing where model performance is compared with random recommendation, analyzing key performance evaluators such as CTR, Sequence quality, top sequences, length comparison etcProject B- Developed a Regression model predicting shipment charges for a client dealing in logistics (Vehicle -Transportation in US)- Used Google’s Distance Matrix API and some data scraping methodologies for data completion- Experimented with different Tree based algorithms - DT, RF, XGBoost etc. to encompass key business aspects- Model predicts expected shipment charges after analyzing various key parameters such as - distance, vehicle type, date of shipment etc.Project A- Developed a Blink Detection model for a client who specializes in professional photography- Conceptualized and Developed a cascaded automated data pipeline which filters images where a human eye blink is detected.- Used MTCNN for Face detection, DLIB for eye detection and hypertune for different pre-trained architectures - MobileNet, DenseNet, ResNet- Cascaded pipeline performs Face detection, Eye extraction and Blink detection on input batch of images and generates an output json which works as input for future stages. Show less

• 

  Tiger Analytics

  Aug 2021 - now

  • Data Scientist

    Aug 2023 - now

  • Senior Analyst - Data Science

    Aug 2021 - Jul 2023