Neeha Girja

Open to work

Software Engineer

I develop scalable systems. I develop solutions that ship fast and perform under pressure.

Scroll

About

EnterpriseFull-stackCloudAI/MLAgile

Hey! I’m Neeha, a software engineer who loves building scalable web apps and end-to-end solutions. At Ernst & Young (EY), I worked on enterprise projects focused on efficiency, security, and performance, with a strong foundation in API design, system integration, and agile delivery.

I’m currently pursuing a Master’s in Software Engineering Systems at Northeastern University, where I’m deepening my skills in full-stack development, cloud architectures, and modern software practices. I enjoy using the right tools to build reliable, scalable, and user-first products.

Selected Work

Uber-Style Ride-Hailing Database Model

Ride-hailing platforms need a database that captures customers, drivers, trips, payments, surge pricing, and real-time status while enforcing role-based access for different user types.

Architecture

Normalized schema: CUSTOMER, DRIVER, DRIVER_STATUS, TRIP, TRIP_STATUS, PAYMENT, VEHICLE, RIDE_TYPE, STATUS, STATUS_MASTER, SURGE_PRICING
Role-based access: ADMIN (bootstrap), APPLICATION_ADMIN, CUSTOMER, DRIVER, FLEET_MANAGER with grants per role
Views for filtered access: customer profile/trips, driver trips/earnings, support cases, low-rated trips, drivers without license
Stored procedures and DML for sample data and reproducible setup

Tech Stack

Oracle DatabaseSQL (DDL, DML)PL/SQLRoles and grants

Scale

Multi-role production-style schema with views and procedures

Challenges

Designed schema to support trip lifecycle, surge pricing, and driver availability
Implemented 8+ views for role-based and analytical access without exposing raw tables
Documented ERD, run order, and grant strategy for clean deployment

Impact

Delivers a reusable database model for ride-hailing with clear separation of concerns and secure, auditable access for each user type.

Code

FixFinder — Service Marketplace

Users need a single place to find and book local service providers (electricians, handymen, etc.), while providers need a way to list services, set pricing, and track bookings and revenue.

Architecture

Full-stack app: React frontend (app/) and Node/Express REST API (service/) with MongoDB
Domain-driven design: User, ServiceProvider, Service, Booking, Payment with clear boundaries
End-user flows: search by location/requirements, view profiles, book services, manage bookings in a dashboard
Provider flows: manage offerings and pricing, dashboard for revenue and performance

Tech Stack

ReactReduxNode.jsExpress.jsMongoDBSCSSPWAInternationalization

Scale

Web-based platform with CRUD, routing, and global state

Challenges

REST API designed with DDD and Microsoft API design best practices
Modular UI with reusable components and professional UX
PWA and i18n for offline use and multilingual support

Impact

Connects end users with local service providers through search, booking, and dashboards; streamlines discovery and management for both sides.

Code

Cloud-Native Web Application

A full cloud application needs a backend API (users, products), infrastructure provisioning, and serverless email verification—all with CI/CD, multi-environment support, and secure, repeatable deployment.

Architecture

Web app (Node/Express, Sequelize, PostgreSQL): REST API /healthz, /v1/user, /v1/product; Packer builds immutable Ubuntu AMI with Node 18, systemd; 54 integration tests
Infrastructure (Terraform): VPC, ASG, ALB, RDS, S3, SNS; Lambda + DynamoDB for email verification; KMS, IAM, Route53; Dev/Demo environments
Serverless (Lambda): Consumes SNS on user registration; sends verification emails via SendGrid; deduplication with DynamoDB; packaged and deployed via GitHub Actions
CI/CD: Packer check/build for AMI; Terraform format/validate/tfsec; Lambda lint/bundle/deploy to Dev then Demo

Tech Stack

Node.jsExpressSequelizePostgreSQLTerraformPackerAWS (EC2, ASG, ALB, RDS, S3, SNS, Lambda, DynamoDB, Route53)GitHub Actions

Scale

Multi-environment (Dev/Demo); ALB + ASG (min 3, max 5); Lambda email verification

Challenges

Integrating webapp, infra, and serverless repos into one deployment pipeline
Packer AMI build + Terraform provisioning + Lambda deploy across accounts
Health checks, auth, validation, and SNS→Lambda→SendGrid flow with dedup

Impact

Delivers a production-ready cloud-native API with automated infra, AMI builds, and serverless email verification across Dev and Demo environments.

Code

Experience

DMSB AI Strategic Hub at Northeastern University

Sep 2025 – Present

Research Assistant

Boston, MA

Architected production LLM inference infrastructure (vLLM, KV cache, continuous batching) achieving 6x throughput on GPU clusters. Led AI applications with RAG pipelines serving 1000+ professors; Docker, PM2, MCP integrations, nginx, and automated deployments.

Ernst & Young

Sep 2022 – Apr 2024

Software Engineer

Bangalore, India

Built real-time credit scoring microservice (Kafka, Spring Boot) processing 100K+ transactions daily for a European banking client. Developed 100+ REST APIs, OAuth 2.0, Redis caching; microservices architecture and CI/CD with Jenkins and Docker; mentored 15+ trainees.

Ernst & Young

Jun 2021 – Aug 2022

Associate Software Engineer

Bangalore, India

Implemented enterprise communication solutions for 5+ global clients; managed call server databases for 1M+ daily transactions. Resolved 100+ critical issues monthly, 99% uptime, reduced response time from 2h to 30min.

JavaScriptJavaPythonC/C++MatlabHTMLCSSTypeScriptREST APIPWASwaggerReactAngularNode.js.NET CoreSpring BootSequelizeMUIBootstrapJavaScriptJavaPythonC/C++MatlabHTMLCSSTypeScriptREST APIPWASwaggerReactAngularNode.js.NET CoreSpring BootSequelizeMUIBootstrap

Technologies and Tools I work with

(and many more)

OracleMS SQLMongoDBPostgreSQLRedisAWSAzureDockerLinuxGitGitHubJiraSketchMS OfficeVS CodeNotionSlackFigmaOracleMS SQLMongoDBPostgreSQLRedisAWSAzureDockerLinuxGitGitHubJiraSketchMS OfficeVS CodeNotionSlackFigma