Sameer Korlahalli

Work

	Software Engineer Lutron Electronics Working on revamping a Customer Facing Log Tool to work with multiple processors for campus-wide lighting control systems Worked on integration of new IoT devices with existing web services using .NET (C#) and Backbone.js
	Software Development Intern NYU IT Services Public Surveillance System (Python, Kafka, JavaScript, Raspberry Pi) Developing a real-time surveillance system for university infrastructure Learning Analytics (Python, AWS & pySpark) Developing data integration and streaming services from Oracle DB and video streaming APIs using AWS Lambda, S3, RDS Specifically working on improving the scalability of the integration and processing of incoming data using pySpark API Approval Request System (Python, AngularJS & AWS) Developed an API approval request system for data and related APIs pertaining to all departments and schools under the university Used MySQL in RDS and S3 as a secondary data store for request related information and Lambda functions as the back end
	Full Stack Developer Intern Big Apple Buddy Developed a critical module for significantly decreasing the response time to consumer product quotation requests, thereby increasing the conversion rate of the leads Worked on the design, implementation, deployment as well as its seamless integration with the company’s existing workflow using Django, PostgreSQL, AWS along with Zendesk, FedEx APIs

Developed a social networking application using Distributed systems concepts and utilizing a multi-server architecture using Golang

Developed an application to recognize food items from an uploaded image and list nearby restaurants offering the identified food item

Used TensorFlow to retrain the penultimate layer of Inception V3 model on a food categories dataset with an accuracy of 90.4%
Utilized S3 to host a static webpage and used Lambda Functions for business logic
Deployed the trained model on EC2, from which results for a particular image were fetched by a lambda function
Implemented cognito for user athentication and RDS for storing restaurant information

Designed and implemented a dynamic web application to capture live tweets using the Twitter API and visualize them on a map

Used SQS to queue the incoming live tweets and extracted the sentiment of each tweet using Watson NLU API
Used Elasticsearch to store the processed tweets
Implemented the application as a serverless system using Lambda functions and Amazon’s API gateway

Developed a project as part of the Audio Content Analysis course to detect arrhythmia from PCG signals

Used signal processing concepts such as Onset Detection for pre-processing and Spectrograms for feature extraction
Developed a 3 layer Convolutional Neural Net for the detection of arrhythmia using Keras (Python)
Trained the model using High Performance Cluster Computing (HPC)

Worked on a project to predict mental state of students using smartphone usage data such as sensory data, call logs & app usage

Preprocessed the data by merging discrete datasets and eliminated redundancies using dimensionality reduction
Used Lasso along with linear regression to identify the relevant features and predict stress levels of students