Jul 2019 – Dec 2022 · 3 yrs 5 mos · Bengaluru, Karnataka, India · Remote

• We explored, analyzed, visualized data and built ML models for various sectors.
• Automatic customer complaints classification project:
• This project classifies customer complaints to relevant topics based on following approaches:
• ➢ For a leading financial company in Bangalore, the clients’ complaints dataset is explored, and Topic modelling is done using NMF (Non-negative Matrix Factorization) to get the relevant topics of complaints.
• ➢ Later, based on the topics that are mapped to the customer complaints, we used a supervised model to predict the relevant topic of any new incoming complaints.
• Complete End to End project on Machine learning operations (MLOps) using Jarvislabs:
• To reduce CAC (Customer Acquisition Cost) of an EdTech company ‘CodePro’, we calculated Lead scores to determine whether a lead will opt for the course or not ultimately.
• Gesture recognition project:
• This project classifies the gestures correctly based on the given input video.
• ➢A cool feature that is developed in the smart-TV can recognize five different gestures performed by the user which will help users control the TV without using a remote.
• ➢3 kinds of architecture used to solve this problem are:
• 1. 3D Convolutional network
• 2. CNN + RNN architecture
• 3. Transfer learning with GRU (Gated recurrent units)
• Telecom Churn prediction project:
• This project predicts whether a client tends to churn out or not, given the client details.
• ➢A leading telecom company’s client dataset is explored and various models like Logistic Regression, Logistic Regression with PCA, Random Forest technique, PCA with Random Forest, XGBoost to improve accuracy with tuned hyper parameters are built on top of it.
• ➢These models are compared and the one that works well on unseen data is taken as the final model among the classes of models.
• Analyzing Loan Dataset of a lending finance company:
• This project is focused to figure out whether a given client with basic details tends to default or repay the loan promptly.

Jun 2014 – Jul 2015 · 1 yr 1 mo · Chennai, Tamil Nadu, India · On-site

• Ensured software quality by performing regression and system-level testing before release. Done manual testing for an insurance project (Liberty Mutual Insurance Company)

Jan 2014 – Jun 2014 · 5 mos · Bengaluru, Karnataka, India · On-site

• IP Traceback Scheme using mathematical derivation for software exploit attacks:
• Implementation of IP traceback scheme for DDoS (Distributed Denial of Service) attacks with the parameters like Degree of Routers, Upper Interface and Hop counts.
• Reliable IP Traceback scheme by means of k-means clustering for flood attacks:
• DDoS attacks usually deal with the flooded packets which utilize the services. Using the count of flooded packets, the upstream routers are computed.