Kunal AI Portfolio

Built using data from my life.

What I Do

I craft modern, efficient, and user-friendly applications that solve real-world problems. With expertise spanning both frontend and backend technologies, I deliver complete solutions that combine beautiful interfaces with powerful functionality.

Backend Development

Building robust, scalable, and secure server-side applications with modern architecture patterns

AI & Machine Learning

Building intelligent applications with LLMs, RAG systems, and ML models to solve complex problems

API Development

Designing and implementing RESTful and GraphQL APIs with a focus on performance and security

Frontend Development

Creating responsive, interactive, and intuitive user interfaces with modern frameworks

Work Experience

Founding Software Engineer

VibeSea AI

Middletown, DEAug 2025 - Present

Engineered core backend microservices (TypeScript, Python/FastAPI) and an AI data pipeline for a next-generation anti-cheating desktop app, enabling real-time analysis of user screenshots and system logs.

Led a team of 4 engineers in designing and building the mission-critical backend infrastructure for an anti-cheating app; scaled the initial system from zero to support 3000+ active users (candidates, interviewers, HR) across 20 companies.

Architected and deployed 4 scalable, event-driven microservices (Python/FastAPI, TypeScript) to manage authentication, interview lifecycle, and notifications, ensuring 99% uptime for 1,000+ concurrent sessions.

Engineered the high-throughput, real-time data pipeline for the core anti-cheating engine, processing 50 screenshots/min and 500 MB of OS logs per session, which was a critical component in securing $500k Seed funding.

Engineered the AI/LLM integration to automatically analyze visual and log data, generating violation summaries with 97% accuracy and reducing manual review time by 70%, directly improving operational efficiency.

Collaborated in a fast-paced, cross-functional startup environment to define system architecture, refine product features, and ensure seamless integration between the backend services and the React.js frontend.

Developed the cross-platform Electron.js app for candidates to enable screen capture, log collection, and low-latency data transfer.

Software Engineer - Intern

USC FPM

Los Angeles, CAJan 2024 - April 2025

Contributed to the development of a campus-wide inspection management web application for USC FPM, streamlining door data handling, offline operations, and compliance workflows.

Led a team of 7 developers in building a full-stack campus-wide asset management system (React.js, Node.js, MongoDB) for USC inspection officers, with interactive floor maps, reducing inspection reporting time by 35% and cutting asset lookup latency by 60%.

Built version-controlled data logging for every door update, allowing users to view history of photos, annotations, and changes over time. Engineered critical field features, using IndexedDB including offline-first support (slashing delays by 40%) and QR code/photo annotation (accelerating documentation by 50%) and one-click generation of weekly, monthly, and annual reports.

Created a secure upload mechanism for inspection images and documents, supporting both drag-and-drop and in-app camera usage. Implemented enterprise-grade security using SAML 2.0 for USC Shibboleth SSO and JWT-secured APIs for role-based access control.

Constructed an intuitive web application based dashboard for team leads to monitor inspection progress, sync status, and pending tasks across buildings.

Added support for customizable tags and filters, helping inspectors quickly categorize and search large volumes of door records. Automated API (Postman, REST Assured) and UI (selenium) testing to raise test coverage to 85% and cut regression bugs by 45%.

Maintained audit logs of every change made by users, supporting compliance tracking and operational transparency.

Software Engineer

Bank of America

Mumbai, IndiaAug 2022 - Aug 2023

Developed and optimized backend microservices for Bank of America's proprietary cloud platform, enhancing scalability, security, and efficiency, and played a key role in shaping the bank's cloud infrastructure strategy.

Contributed to the development of Bank of America's proprietary cloud platform, building and optimizing backend microservices using Java and Spring Boot.

Designed and implemented scalable services for request routing, SSL termination, and API rate limiting, reducing latency by 30% and improving service reliability.

Developed dynamic load-balancing and secure API exposure mechanisms, increasing request-handling capacity by 25% and reducing system downtime by 20%.

Collaborated with cross-functional teams to integrate Kubernetes for containerized service deployment, improving system scalability and resilience. Acted as a key technical liaison in a fast-paced Agile/Scrum environment, collaborating daily with QA teams to define test strategies and translating complex stakeholder requirements into actionable user stories.

Recognized with the Golden Award by Senior Vice President of Digital Solutions team for resolving critical backend optimization of a core-system bottleneck in the real-time event stream in the BACNet 23.1.2 release, enhancing the digital banking experience and boosting message throughput by 40%.

AI / ML Intern

Defence Research & Development Organization (DRDO)

Pune, IndiaJun 2021 - Oct 2021

Engineered an AI-driven Electronic Countermeasure (ECM) system using CNNs, LSTMs, and Q-learning to optimize real-time jamming. Integrated models into an ECM pod, successfully increasing pilot situational awareness by 40% and overall system capabilities by 27%.

Engineered a hybrid deep learning model using CNNs for spectrogram pattern recognition and RNNs (LSTM) for temporal signal evaluation, optimizing real-time jamming strategies.

Achieved a 40% improvement in pilot situational awareness and countermeasure decision-making by successfully deploying the CNN/LSTM model.

Developed a reinforcement learning agent using Q-learning and Deep Q-Networks (DQN) to determine optimal jamming actions in dynamic signal environments.

Integrated all neural network models and jamming algorithms into a cohesive Electronic Countermeasure (ECM) pod system, enhancing operational capabilities by 27%.

Education

Master of Science in Computer Science

University of Southern California (USC)

Los Angeles, CAAug 2023 - May 2025

GPA: 3.62 / 4.0

Relevant Coursework

CSCI 571 : Web Technologies

CSCI 585 : Databases

CSCI 566 : Deep Learning

CSCI 544 : Natural Language Processing

CSCI 572 : Information Retrieval & RAG

CSCI 570 : Algorithms

CSCI 577B : Software Engineering

Bachelor of Technology (BTech) in Computer Engineering

University of Mumbai

Mumbai, IndiaAug 2018 - Aug 2022

Won 1st place in Smart India Hackathon (SIH) 2021. Led a team of 5 to develop an innovative platform for immigrants in metropolitan areas, simplifying access to healthcare facilities.

What I Do

Backend Development

AI & Machine Learning

API Development

Frontend Development

Work Experience

Founding Software Engineer

Software Engineer - Intern

Software Engineer

AI / ML Intern

Education

Master of Science in Computer Science

Relevant Coursework

Bachelor of Technology (BTech) in Computer Engineering

Featured Projects

ShopSphere - E-Commerce Microservice Platform

Tech Stack

Multi-Agent AI Travel Advisor

Tech Stack

StockVortex - Stock Market Simulator

Tech Stack

FinGPT-Advisor – LLM Fine-Tuning for Financial Chatbot

Tech Stack

EduNexus - Full-Stack Learning Management System

Tech Stack

CodeCollab - Online Code Editor IDE

Tech Stack

CareerVista - Online Job Portal

Tech Stack

QuickScan - AI-powered Document Scanner

Tech Stack

Technologies & Tools

Spring Boot

React

Next.js

TypeScript

JavaScript

Node.js

Python

Java

HTML5

CSS3

Tailwind CSS

MongoDB

PostgreSQL

Docker

AWS

Git

SQL

TensorFlow

PyTorch

OpenAI

LangChain

Scikit-learn

Keras

Jupyter

Get in Touch

Email

Phone

Location

Connect on Social Media