Backend Engineer • AI Research • Computer Vision • Full Stack

I enjoy building systems that solve real engineering problems.

From AI evaluation frameworks and privacy-preserving interfaces to computer vision applications, my focus is on designing software that is scalable, measurable, and practical.

Currently working as an AI Research Intern at Appy Pie, where I evaluate multimodal and agentic AI systems across text, image, audio, video, and 3D domains.

AI Evaluation • Python • Three.js • Docker

A model-agnostic evaluation framework for Image-to-3D models.

• Statistical Bradley–Terry / ELO ranking
• Automated mesh-quality analysis
• Blind human & VLM voting
• O(n²) → O(n) evaluation optimization
• Dockerized Three.js comparison platform

Privacy-Preserving AI Interface

Protects sensitive information before prompts reach external LLMs.

• Regex + Named Entity Recognition
• Browser Extension
• Live ChatGPT protection
• Privacy-first middleware

Computer Vision Platform

Image processing platform built using FastAPI + React.

Features:

• Image Enhancement
• Histogram Equalization
• Gaussian Denoising
• Bicubic Upscaling
• C++ Steganography Engine
• OpenCV Processing Pipeline

Working on:

• Multimodal AI Benchmarking
• Agentic AI Evaluation
• Dataset Creation
• Product Research
• Failure Mode Analysis

Python • C++ • Go • JavaScript • SQL • C

FastAPI • REST APIs • PostgreSQL • MySQL • SQLite

React • HTML • CSS • Three.js

OpenCV • NumPy • SciPy • Pillow • PyMeshLab • trimesh

Docker • Git • Linux • VS Code • Figma

🏆 5× Hackathon Podium Finisher

🥈 Build with TRAE — NIT Delhi

🥈 HackMSIT 1.0

🥈 Mind Installers 4.0

🥉 Hackverse 2.0

🥉 IGNITE 1.0

"Build systems that outlive the code."