Available for SDE & AI Roles

Gaurav Kumar
Verma

Software Developer | AI/ML Systems

I build AI-powered backend systems and full-stack applications focused on real-world impact, privacy, and scalability.

View Projects GitHub Profile
98.66% Accuracy Healthcare AI Model
Secure AI Pipeline Local Agent Execution
ML Analytics Pollution Forecasting

// 01 — Executive Summary

Bridging the gap between
data models and user experiences.

I am a Software Engineer specializing in Full-Stack Development and AI/ML integrations. I build secure, scalable systems—from hybrid-intent AI pipelines to real-time geospatial dashboards.

Currently completing my B.Tech in CSE at Jaypee Institute of Information Technology (CGPA: 7.9). I excel in high-pressure environments, frequently competing in national-level hackathons.

5+
Full-Stack Projects
Oracle
Certified AI Associate
3+
Hackathon Awards

// 02 — Technical Arsenal

Core Competencies

⚙️
Backend & Systems
Python FastAPI C++ SQL Firebase Git
🧠
AI & Machine Learning
Scikit-learn Pandas NumPy ChromaDB Computer Vision
💻
Frontend & UI
React Next.js TypeScript Tailwind CSS

// 03 — Engineering Proof of Work

Featured Projects

Deep Learning • 2026

Brain Tumor Classification

Stack: Python, PyTorch/TensorFlow, Computer Vision Problem

Need for high-accuracy automated diagnosis in MRI scans to assist radiologists and reduce diagnostic latency.

Solution

Built a deep learning classification model using EfficientNetV2-S architecture and implemented Grad-CAM for model interpretability (visualizing diagnostic regions).

Impact: Achieved 98.66% accuracy and 0.9995 ROC-AUC on standard MRI datasets, providing clinical-grade performance.
GitHub
AI & Automation • 2026

ATLAS: Local AI Desktop Agent

Stack: Python, FastAPI, ChromaDB, Ollama Problem

Cloud AI assistants introduce privacy risks and high latency for local system operations and file management.

Solution

Engineered a local-first assistant with a hybrid intent engine (rule-based + LLM) and RAG-based context memory using ChromaDB.

Impact: Achieved ~85% intent classification accuracy and processed 100+ simulated inputs with zero external data exposure.
GitHub
Platform Engineering • 2026

CodeVibe: AI Prompt Hub

Stack: Next.js, TypeScript, Firebase Problem

Developers need a secure way to manage AI prompts without exposing API keys to third-party server backends.

Solution

Architected a BYOK (Bring Your Own Key) platform with client-side API execution and Firebase for secure state management.

Impact: Reduced API latency by ~30% via client-side routing and secured 100% of user keys through zero-server-storage architecture.
GitHub Live Demo
Data Intelligence • 2026

Pollution Analytics Platform

Stack: React, FastAPI, Scikit-learn, GeoJSON Problem

Lack of ward-level environmental data prevented accurate personal health decisions and civic policy planning.

Solution

Developed a geospatial monitoring portal with Scikit-learn models to forecast pollution spikes at the neighborhood level.

Impact: Processed 10,000+ data points to achieve ~92% accuracy in predicting ward-level AQI spikes, enabling preventive measures.
GitHub

// 04 — Recognition

Awards & Certifications

🏆
Oracle Certified Associate
Cloud Infrastructure & AI Foundations 2025
🥇
Hackoverflow 9.0
Certificate of Excellence (Top Performers)
🏅
National Hackathons
Honors in Paranox 2.0 & HackVision

// 05 — Collaboration

Let's build the future.

I am seeking Software Engineering and Data/AI roles where I can apply technical rigour to high-impact problems.

gaurav.k.verma23@gmail.com Download Resume (PDF)