● Trained and deployed an image classification model to display the best image on listings pages using transfer learning which improved the click through rate● Developed a process to discover new business opportunities from ‘search query text’ using topic modeling (LDA) technique● Developed a machine learning pipeline for predicting booking cancellation and revenue● Deployed a machine learning pipeline for predicting property score based on historical data● Worked on the full machine learning lifecycle including cleaning, preprocessing, feature engineering, model training and deployment● Set up Amazon EC2 for developing machine learning models and deployed using Docker● Created a Flask based web-interface for in-house data visualization, reporting and update● Experienced in working with external consultants, cross-functional team located home and abroad in an agile environment● Technology Stack: Python, TensorFlow, scikit-learn, XGBoost, pandas, numpy, MLflow, NLTK, spaCy, gensim, PostgreSQL, Flask, Docker, Tableau

Courses:● Introduction to Computer Systems● Data Communications/Networking

● Worked in a fast paced Agile development environment● Developed a machine learning pipeline for human activity recognition from sensor data (smartphone sensors and wrist-worn accelerometers) using Keras and Tensorflow● Re-implemented existing algorithm in PySpark for faster (3x) sensor data processing● Worked on a team for developing a data collection, processing, storage, analytics and visualization pipeline using MySQL, MongoDB, InfluxDB, Node.js, D3.js, Chart.js

● Developed Python script for a Raspberry Pi controlled parking automation system● Developed an android app that can control a Raspberry Pi using Bluetooth API

Courses Instructed:● Introduction to C/C++ programming language         ● Assembly language and µ-processor design● Signals and systems● Electrical circuits II (AC circuits)

Responsibilities:● Network design and optimization for emerging heterogeneous network ● Algorithm design and implementation for smart homes with heterogeneous network● Applying Convex optimization and Game theoretical concepts for network optimization

Accomplishments:● Proposed a novel algorithm based on IMU (inertial measurement unit) data and positioning information from a novel angle of arrival (AOA) sensor. ● Used and modified Extended Kalman filtereing (EKF) for data integration for the proposed algorithm. The modified algorithm provided 60% improvement in the best case and at least 7% improvement in the worst case scenario. ● Implemented the proposed algorithm MATLAB and C.

Responsibilities:(http://ppfhinitiative.com/Students.html)● In lieu with the common goal and vision of the ICICS and TELUS People and Planet Friendly Home project, design and optimize network protocols for emerging smart homes.              ● Design efficient resource allocation and scheduling technique for HetNets.