Headline

AI/ML Engineer | Building Full-Stack AI Systems with LLMs, NLP & Computer Vision | React · FastAPI · Docker · Kubernetes

About

I realized early that building models is easy. Building systems that work in the real world is not. During my AI internship at Infosys, I built SynnccIT, an AI-powered cloud IDE — a React/TypeScript frontend backed by a distributed FastAPI microservice architecture, with a built-in AI coding assistant powered by Google Gemini, a real-time WebSocket terminal, and an in-browser file explorer with native file-system access. That experience taught me the real challenge in AI isn't accuracy alone, but scalability, reliability, and integration. Since then I've kept building end-to-end AI systems across the stack — from ML models to production-ready APIs to the frontends people actually use: What I've built: 🔧 SynnccIT — AI-powered cloud IDE with a distributed FastAPI backend and Gemini-powered coding assistant (Infosys internship) 📄 ATS Resume Scorer — full-stack app scoring resume-JD compatibility via semantic similarity (Sentence Transformers + spaCy), with Groq Llama 3 for personalized feedback, Supabase auth, and PDF report generation ☸️ AI-Powered Kubernetes Platform — ML-based failure prediction for K8s clusters, containerized and orchestrated with Docker/Kubernetes, with automated recovery workflows 🏋️ AI Gym Coach — real-time pose estimation (MediaPipe/OpenCV) for form analysis across 5 exercise types, with automatic rep counting and Groq-powered voice coaching via gTTS 📝 Published two papers on ML for fake news and SMS spam detection (IJRASET, 2026) I care about clean architecture, reproducibility, and writing code that's maintainable and production-aware. I don't build academic projects — I build systems with engineering depth, from model to deployment. Open to AI/ML engineering roles — internship or full-time — where I can keep working on real production systems, not just notebooks.

Location

Bengaluru

AI stack

claude code

codex

github copilot

lovable

notebooklm

Work experience

AI Intern at Infosys Springboard, Dec 2025 - Feb 2026 · 3 mos, India, React, TypeScript, Tailwind CSS, Python, FastAPI, Google Gemini Built SynnccIT, an AI-powered cloud IDE, pairing a React/TypeScript frontend with a distributed FastAPI microservice backend handling file management, terminal access, testing, and monitoring Integrated a built-in AI coding assistant using Google Gemini for code generation, terminal command translation, and task planning Implemented a real-time, WebSocket-based cross-platform terminal and an in-browser file explorer with native file-system access Designed dedicated backend services for automated test execution, simulation, and flowcharting Links: GitHub: https://github.com/KushiArun/Development-of-AI-Driven-Code-Reviewer · Live Demo: https://synncc-it.vercel.app/

Data Science Intern at Unified Mentor Private Limited, Aug 2025 - Nov 2025 · 3 mo, India, Built predictive ML models following exploratory data analysis and feature engineering on structured datasets Evaluated model performance using accuracy, precision, recall, and cross-validation metrics Developed visualization dashboards to communicate analytical insights to stakeholders

Education

B.E, Artificial Intelligence & Data Science, East west institute of technology

PUC, PCME, Vvs Sardar Patel PU College

Public links

github: https://github.com/KushiArun

leetcode: https://leetcode.com/kushiarun

kushi-portfolio-glow.lovable.app: https://kushi-portfolio-glow.lovable.app/