Nikhil Kotte

AI/ML Engineer | Computer Vision, LLM Agents & Multi-Agent Systems | CSE Undergraduate at CBIT

Hyderabad, India

nikhilkotte24@gmail.com

About

I'm a Computer Science Engineering undergraduate at Chaitanya Bharathi Institute of Technology (CBIT), Hyderabad, with a strong interest in Artificial Intelligence, Machine Learning, and building practical software solutions. My experience spans computer vision, AI-assisted development, multi-agent systems, and full-stack applications. As an Artificial Intelligence Intern at Nemetschek Group, I researched and benchmarked AI coding workflows, developing custom agent skills to improve code quality while reducing development costs. During my research internship at IIITH I-Hub Data, I developed a real-time waste classification system using YOLOv11 and transfer learning, achieving 97.83% mAP@50. I also completed an intensive AI/ML training program at IIIT Hyderabad, gaining hands-on experience implementing machine learning algorithms and neural networks. Beyond internships, I enjoy building products and competing in hackathons. My team secured 2nd Prize at Forge Inspira Hackathon 2026, IIT Hyderabad, with Market.Intel, an AI-powered multi-agent market intelligence platform. I've also contributed to ClaimIQ, a full-stack AI-powered insurance claims platform. I'm particularly interested in developing intelligent systems that combine strong software engineering with modern AI capabilities. I enjoy exploring emerging technologies, solving challenging problems, and turning ideas into functional, meaningful applications.

Skills

Python

C++

Java

JavaScript

Ultralytics YOLO

OpenCV

NumPy

pandas

Matplotlib

Supabase

GitHub

GitHub Copilot

Machine Learning

Deep Learning

Computer Vision

Object Detection

Transfer Learning

Supervised Learning

Unsupervised Learning

Model Evaluation

LLM Agents

Prompt Engineering

Multi-Agent Systems

Claude Code

Gemini

Experience

Artificial Intelligence Intern

Nemetschek India Private Limited

Jun 2026 to Aug 2026

Researched and evaluated AI-assisted software development workflows to improve code quality and reduce computational costs. Benchmarked direct prompting, specification-driven development, and custom agent skills on real-world feature implementations. Designed and evaluated repository-specific AI agent skills using metrics including linting compliance, bug-free output, runtime complexity, iteration count, and rework. Demonstrated that custom agent skills could achieve comparable code quality to more expensive workflows at substantially lower credit costs.

Research Intern

iHub-Data, IIIT Hyderabad

Oct 2025 to Dec 2025

Developed a real-time waste classification system using transfer learning with YOLOv11s, capable of categorizing organic, recyclable, biomedical, and miscellaneous waste. Trained the model on 649 labeled images, achieving 96.45% precision, 97.65% recall, and 97.83% mAP@50. Built an end-to-end Python inference pipeline using Ultralytics YOLO, OpenCV, and NumPy, supporting image, video, and live camera detection, object counting, and FPS monitoring.

Projects

Market.Intel — AI-Powered Market Intelligence Platform

Developed a full-stack AI-powered market intelligence platform using a modular multi-agent architecture to automate market research, intelligence gathering, and workflow orchestration. Built Python backend services for autonomous task execution and agent coordination, integrated with a JavaScript dashboard for real-time insights. Won 2nd Prize at Forge Inspira Hackathon 2026, IIT Hyderabad.

Python

JavaScript

Multi-Agent Systems

AI Agents

Full-Stack Development

Real-Time Waste Classification Using YOLOv11

Developed a real-time waste classification system using YOLOv11s and transfer learning to detect and classify organic, recyclable, biomedical, and miscellaneous waste. Trained on a custom dataset of 649 labeled images, achieving 96.45% precision, 97.65% recall, and 97.83% mAP@50. Built an end-to-end Python inference pipeline supporting image, video, and live camera detection, object counting, and FPS monitoring.

Python

YOLOv11

Ultralytics

OpenCV

NumPy

Computer Vision

Deep Learning

Transfer Learning

Object Detection

ClaimIQ — AI-Powered Insurance Claims Platform

Developed a full-stack AI-powered insurance claims platform that analyzes vehicle damage images to generate instant pre-approval recommendations. Built a FastAPI backend integrating YOLO-based damage detection, severity scoring, vision-language explanations, and fraud risk analysis. Used Supabase for authentication and data persistence, with a React and TypeScript dashboard. Ranked among the Top 100 of 620 teams at the KLH University Hackathon.

Python

FastAPI

YOLO

React

TypeScript

Supabase

Computer Vision

Machine Learning

AI/ML Implementations — Machine Learning From Scratch

Implemented core machine learning and deep learning algorithms from scratch as part of the IIITH I-Hub AI/ML program. Developed K-Nearest Neighbors (KNN), perceptrons with gradient descent, forward propagation, backpropagation, and Convolutional Neural Networks (CNNs) using NumPy and PyTorch. Achieved over 97% test accuracy on the MNIST handwritten digit classification dataset.

Python

NumPy

PyTorch

Machine Learning

Deep Learning

Neural Networks

CNN

Gradient Descent

Education

B.E. in Computer Science and Engineering

Chaitanya Bharathi Institute of Technology · 9.32

Sept 2024 to Jun 2028