Aryan Rastogi

Machine Learning & AI Engineer

Jaipur, India

About Me

I train models to learn & they teach me patience.

I am a computer science student working on machine learning, computer vision, NLP, and applied intelligent systems. I craft effective prompts, build clean frontends, and work comfortably across backend systems.

Off screen, I like to slow down, eat well, and catch up on sleep.

Tech Stack

Languages

PythonCppSQLTypescript

ML/AI Frameworks

PyTorchTensorFlowScikit-learnHugging FaceLangChainOpenCV

Full Stack

ReactNext.jsNode.jsREST APIsSupabase

Tools & Platforms

FastAPIAWSGCPDockerKaggle

GitHub Activity

Loading GitHub activity...

LeetCode Stats

Work Experience

SDE Intern

Southern Federal University, ICTIS • Russia (remote)

Developed a web application as an international interdisciplinary project in collaboration with Southern Federal University (ICTIS) & MUJ to support the adaptation of international students. The platform acts as a digital assistant that helps to explore, understand and navigate the campus facilities and improves accessibility for new international students.

Oct - Nov 2025

AI Engineer Intern

Elumina Street Solutions • Jaipur

Developed a violence detection model using CNN (MobileNet) and LSTM in TensorFlow for video-based behavior recognition; improved accuracy with focal loss and class weighting. Implemented and evaluated a real-time weapon detection system using YOLOv8 and OpenCV. Collaborated with the AI research team to analyze model performance, optimize hyperparameters, and enhance detection.

April - July 2025

Research Intern

AI²: computAtional linguIstics & vIsion Lab • MUJ

Conducting an ablation study on LLM creativity, evaluating the impact of fine-tuning methods (SFT, LoRA, QLoRA, RLHF, DPO) on creative performance using DAT and AUT benchmarks.

Jan 2025 - Present

Live Projects

perpleX

AI-powered text humanizer that transforms AI-generated content into natural, human-like text.

AINLPReactNext.js

teXlab

Converts natural language and NLP outputs into professional LaTeX diagrams and tables with intelligent formatting.

NLPLaTeXReactNext.js

Hackathon Builds

Roadeye

AI-powered road audit system using deep learning and computer vision to detect road defects and generate professional audit reports.

PythonPyTorchOpenCVFastAPI

Finalist @ National Road Safety Hackathon 2025, IIT Madras

Prompt2Plugin

AI-powered Chrome extension generator that transforms natural language prompts into functional extensions and n8n automation workflows.

ReactTypescriptn8nChrome API

Project @ CodeClash 2025

Uniway

Campus navigation platform featuring interactive maps, smart search, route guidance, and student services with AI-ready extensions.

ReactOSRMGoogle JS APIsLeaflet

Winner @ Uniway Hackathon 2025

Geoverse

AI-powered conversational system that transforms complex satellite data into instant, contextual answers using NLP and knowledge graphs.

PythonNLPKnowledge GraphsFastAPI

Project @ BAH ISRO, 2025

Achievements & Recognition

National Road Safety Hackathon Finalist

National Road Safety Hackathon 2025, IIT Madras

Shortlisted among the top 10 finalist teams nationwide from 3.6k+ participants & presented at the National Road Safety Hackathon organised by COERs, IIT Madras.

UniWay Hackathon Winner

Uniway MUJ Hackathon 2025

Secured 1st place, won a cash prize, and earned an opportunity to collaborate on a project with Southern Federal University, Russia.

Student Excellence Awardee

Manipal University Jaipur, CSE dept 2025

Awarded for outstanding technical performance, hackathon achievements, and project based innovation during the 3rd & 4th semester.

Solana Bounty Winner

Superteam India, 2025

Won a $60 bounty and became a Global Superteam Contributor within the Solana ecosystem.

Dean's List Awardee

Manipal University Jaipur, CSE dept 2025

Named to the Dean's List in the 2nd semester for academic excellence.

Blogs

Implementing Deep Learning by Diving Deep into the Black Box

Medium • Published 2026

An Implementation of deep learning in three stages: building a neural network from scratch using NumPy, predicting customer churn with an ANN, and classifying images with a CNN.

Deep LearningCNNANN

A Cursor and Google Colab Workflow for Machine Learning: Replacing Jupyter and ChatGPT

Medium • Published 2026

Exploring how to leverage Cursor and Google Colab together to create an efficient machine learning workflow that combines the best of both worlds - replacing traditional Jupyter notebooks and ChatGPT interactions with a more integrated development experience.

Machine LearningCursorGoogle ColabWorkflow

Knowledge Base

Deep Learning

Complete collection of deep learning concepts, architectures, and implementations.

Neural NetworksCNNANN

Python Programming

Comprehensive Python programming short notes covering basics to advanced topics.

FunctionsOOPsAlgorithms

NLP

Notes on Natural Language Processing and implementations.

NLPRNNsTransformers

Leetcode

Leetcode solutions understanding and explanations.

DSAcppmath

Live ML Playground

#start session & press shift+enter to execute cell

import time
import sys
import random

WIDTH = 60
HEIGHT = 20

def clear():
    sys.stdout.write("[2J[H")

# Generate data (y = 0.5x + noise), first quadrant only
data = [(x, 0.5*x + random.uniform(-1, 1)) for x in range(1, 21)]

# Initial line y = mx + b
m = random.uniform(0, 1)
b = random.uniform(0, 2)
lr = 0.01

def predict(x):
    return m*x + b

for step in range(5):
    clear()
    print('Linear Regression Learning\n')
    print(f'Step {step+1}/5 | slope={m:.2f} intercept={b:.2f}\n')

for y in range(HEIGHT, -1, -1):
        row = ""
        for x in range(WIDTH + 1):
            px = x / 3
            py = y / 2

char = " "

# Axes (first quadrant)
            if x == 0:
                char = "|"
            if y == 0:
                char = "-"
            if x == 0 and y == 0:
                char = "+"

# Data points
            for dx, dy in data:
                if abs(px - dx) < 0.3 and abs(py - dy) < 0.3:
                    char = "●"

# Regression line
            if abs(py - predict(px)) < 0.3:
                char = "="

row += char
        print(row)

# Gradient descent update
    dm = sum((predict(x) - y)*x for x, y in data) / len(data)
    db = sum(predict(x) - y for x, y in data) / len(data)

m -= lr * dm
    b -= lr * db

time.sleep(0.3)

print("\n...where learning begins!")