Open to work

Hi, I'm Aman.
A software engineer.

Turning computer vision ideas into usable products and developer tools.

Aman Harsh
Full_Stack_Engineer 0.99
Full StackEngineer
3+Years Exp.
Available

Featured Projects

YOLO NDJSON Converter
Open Source
YOLO NDJSON Converter
Open-source desktop app that converts Ultralytics YOLO NDJSON annotation exports into 12+ ML dataset formats (YOLO v5-v12, COCO, Pascal VOC). Features parallel downloads, runs fully offline, and works cross-platform.
Rust
Tauri
React
TypeScript
PawVision
Live
PawVision
A web app that uses AI to analyze your facial features and reveal your true pet personality traits and match percentages. Do you give off cat or dog energy? Upload your photo and discover if you're more feline or canine at heart.
Python
FastAPI
Next.js
React
Pashudhan
Hackathon
Pashudhan
A web app classifying 75 Indian livestock breeds using AI-powered image recognition with multi-platform support (desktop/mobile/tablet). Built for Smart India Hackathon. Upload an image to identify Indian cattle and buffalo breeds with precision.
Python
FastAPI
Next.js
React

Skills

Languages
PythonJavaRustC++CJavaScriptShell
Frontend
Next.jsReactTypeScriptTailwindCSSVite
Backend
FastAPISpring BootNode.jsPostgreSQL
Computer Vision
UltralyticsRoboflowPyTorchOpenCV
DevOps & Tools
GitDockerCloudflareRailwayGoogle CloudGitHub Actions

About Me

Aman Harsh

Aman Harsh

Full Stack Engineer

NIT Trichy, Computer Science
Pune, India
contact@amanharsh.in

I'm Aman Harsh, a Full Stack Software Engineer who builds fast, scalable, end-to-end products.

I specialize in turning ideas into production-ready applications, owning everything from frontend to backend to deployment.

My work focuses on full-stack engineering and Vision AI, where I build and ship real-world AI products, not just models.

Recently, I've built products like YOLO NDJSON Converter, Pashudhan AI, and PawVision AI, bringing computer vision into practical, usable tools.

I care deeply about performance, clean architecture, and shipping software that people actually use.