Tarun Eswar

Work Experience

Software Engineer Intern/Trainee

Rakuten Group, Inc

June 2022 - August 2022 && July 2023 - August 2023

Y1: Worked on the front and backend for 6th sense—a monitoring platform produced as a business to business model for other developers to track performance vitals of their projects.
Y1: Functioned as a full stack developer using vue.js, node.js, and Elastic Cloud to provide better capabilities to local test acceleration and incident management of the web application.
Y2: Leveraged LLMs in building out new capabilities for GenerativeAI in both the front and back-end
Y2: Focused on Cognitive Engine front-end with typescript, redux, and taiwind

Vue.js
Java
Node.js
Elastic Cloud
LLMs
Python
Flask
Git

Machine Learning Intern

AiDA Technologies Pte Ltd

Jul 2022 - February 2023

Helped process thousands of records of client data for latering modeling processes to create insurance claim models.
Discovered methods to more efficiently prepare data for further analysis with large datasets, typically through Python algorithms.
Data pruning and filtration.

Jupyter Notebook
Pandas
Python

Software Developer

Youth Economics Initative

Sept 2022 - present

Aiding the implementation of infrastructure for the largest student-led economics initiative in the United States.
Serving as a full-stack web developer making use of next.js & node.js.
Maintaining and creating new components of the YEI codebase for user support ahead of large-scale events.

Next.js
Node.js
Typescript
Git

Contracted Work

Tufts Financial Group

Tufts University

August 2022 - March 2023

Previously, Tufts Financial Group had a website that was outdated and difficult to navigate. I worked in coordination with executive members of the organization in order to create a new, revised website. Frameworks, layout, and organization of websites were completely redesigned from scracth. Furthermore, SEO was improved allowing for a 17% increase in activity metrics.

CSS
JS
UI/UX Design
OAuth
Git
SEO

TriggerCalc — Timing Post-Trigger Pregnancy Tests

Independent Project

January 2023 - present

With the numerous hormones and medications required during fertility procedures, women struggling with infertility do not go through the traditional conception and pregnancy process. As a result, I worked with Northeastern lecturer Brett Murphy Hunt to create a publically available calculator to help alleivate this issue. Find the calculator here.

React.js
Bootstrap
Git

WPI Dynamics Lab

Independent Project

February 2023 - present

I worked with the WPI Dynamics Lab to create a reactive model demonstration for a novel machine learning algorithm published in the Journal of Cheminformatics found here. The application allows for a direct connection with the .sav file contents produced by the model which then output model predictions live. The application can be found here.

Python
Linux
Apache2
Flask
OpenCV
Chemparse

Highlighted Projects

Modeling T1 Resting-State MRI Variants Using Convolutional Neural Networks in Diagnosis of OCD [2022-2023]

[Source Code] [Arxiv]

Determining the root cause of obsessive-compulsive disorder is highly difficult. Physicians are often unable to differentiate obsessive-compulsive disorder from major depressive disorder (MDD) and schizophrenia (SZD). This lack of differentiation occurs because OCD is often comorbid with SZD and MDD. The overall aim of this project was to design models for each disorder, develop activation heatmaps, and extract regions of interest. Afterwards, an additional aim was to conduct gene expression analysis to determine the involvement of mGluR5, encoded by GRM, in OCD. A flask app was utilized to demonstrate model preformance.

Tensorflow
Novel Model Design
ResNet50
MobileNet
SkLearn
Pandas
GradCam
Python
Node.js
Python Flask
& more

Independent Community Service Project

After the Covid-19 crisis left its main impact on small businesses, I created an initative to help them re-enter the general markets via web services. I worked with over 7 businesses such as the Virginia Healing Garden, Tincan Flower Co., and more. This project provided these businesses with a method to re-enter the market and gain more customers in a post-pandemic era. More recently, the project was featured on NBC 10 on April 7th, 2022 for exceptional community service.

HTML
CSS
JS
Liquid

MAMS SISO: A Digital Sign-in Sign-out System

The Massachusetts Academy of Math & Science relied on a paper-based attendance system. After realizing the amount of paper and time that was wasted with this method, I worked with Charles Tang and David Barsoum to create a digital sign-in/sign-out system known as MAMS SISO. The dashboard can be found here while the physical RFID scanner used to authorize students has been implemented with a CAD case at Mass Academy.