Abhiram Natarajan

Purdue University, West Lafayette, IN

RV College of Engineering, India

GPA: 9.65 / 10.0
Rank: 5 on 214

Software Engineer 1 - Grade 4

• Designed a mechanism to migrate data from legacy Analyzer Web Services in ElasticSearch to MySQL
• Implemented a caching mechanism with Caffeine and a schema management system with Liquibase
• Converted Webex Analyzer to a Kubernetes based model from a VM model resulting in zero downtime during upgrades
• Split Analyzer UI into two separate instances with their own databases and internal redirection
• Transformed an existing monolithic architecture in Cisco Call Manager to a microservice architecture and built an automation tool to package projects using Apache Maven, reduced build time by a factor of 4.
• Was recognized as the Most Valuable Programmer for Jan 2021

Software Engineer (Freelance)

• Built a full stack marketplace for clients and vendors to work together.
• MongoDB Atlas and Firebase were used in data storage
• Launched as https://bitbaza.io

Student Trainee

• Developed a Real time voice based SPAM detection system using Recurrent Neural Networks
• Used a 1D Convolution and a Sentence Similarity to attain accuracy of over 93% on real time data with minimal latency
• The software was integrated on to Samsung Galaxy mobile devices and a research paper on the same was published

Software Engineer Intern - Summer

• Built a resumed download feature for Cisco DNAC using HTTP requests
• Created a wrapper for Kubernetes API on Python
• An automation tool to validate their functionality was also created

Software Development Engineering Intern

• Developed a full stack web application with Django and AngularJS that were deployed on YellowAnt servers.
• Built Lower Level APIs to interface YellowAnt servers with applications like Gmail, Box, AgileCRM using REST based requests and OAuth2

Voice Call Spam Detection

As a Student Trainee, developed a Real Time Phone Call SPAM Detector using LSTMs, Used 1D Convolution and a sentence similarity model that could attain an accuracy of over 93% on real time data with low latency. The software was tested on the Samsung Bixby platform and has since been integrated on Galaxy devices.

Spam Detection over Call Transcript using Deep Learning
Accepted in Future Technologies Conference, 2021, Vancouver. In press.

Offline Signature Verification Mechanism

Developed a novel signature verification model that uses a Greedy Algorithm approach to test their validity. Feature comparison is performed using Siamese networks that allows for one shot learning using a single sample. Augmentations created using graphical transformations

Signature Warping and Greedy Approach Based Offline Signature Verification
Published in the International Journal of Information Technology, Springer Nature.

Vehicle Detection and Parking Management

Developed a computer vision model using Haar Cascades and YOLO that automated vehicle detection in the R.V.C.E campus. Developed a number plate recognition system in order to check validity of student vehicle parking passes within the campus.

An Approach to Real Time Parking Management using Computer Vision
Presented at ICCCV, 2019, South Korea.

Natural Language to SQL Converter

Developed a structured Natural Language to SQL converter on Java. Researched on methodologies used to extract functional groups from text using hypernyms. Used Ngram technique in intent recognition.

ISQNL: interpretable SQL query synthesizer from natural language input
Presented at ICCCV, 2019, South Korea.

Graduate Teaching Assistant

I worked as a graduate teaching assistant for MA 16500, Calculus 1. Weekly recitation classes were conducted by me to help students solve quizzes and well as homework assignments. My duties also included grading papers, homeworks and holding my own office hours, where students would come and get their doubts clarified at.

Classes on Data Structures and Algorithms

Conducted weekly classes on Data Structures and Algorithms including the concepts and applications of Stacks, Queues, Binary Trees, Tries and Linked Lists using the C programming language. Covered algorithms for Graph Traversals with DFS and BFS, recursion and some basics of Dynamic Programming. Explained approaches to solve relevant problems available on Hackerank and LeetCode using these.

Classes on Machine Learning and Computer Vision

Conducted bi-weekly classes over the semester on the basics of machine learning, solving problems like K-means clustering, manually as well as with APIs. Introduced students to pixel operations, performing basic image processing tasks like contrast stretching, color conversions, and object detection using cascades.

Workshop on Profile Building

Conducted a 4 hour workshop on building a good profile, structuring resumé, working on R&D projects and salary negotiations for students of the 5th and 7th semester sitting for the campus placement and internship drive.