AI ENGINEERING / FULL-STACK / INDORE

.

AI Engineer & Full-Stack Developer

I’m a curious and driven developer who enjoys learning, building, and turning ideas into practical solutions. I’m focused on creating real-world products while continuously growing as an engineer.

EXPLORE PROJECTS ↗LET’S CONNECT

GET TO KNOW ME

Learn.
Build.
Evolve.

FOCUSGrowing toward AI engineering while building intelligent systems and full-stack products.
BUILDBuilding full-stack AI applications across backend systems, APIs, AI workflows, and user interfaces.
EDUCATIONB.Tech in Computer Science & Artificial Intelligence · Medi-Caps University · 2023–2027 · CGPA 7.57/10
OUTSIDE CODESketch artist, guitarist, and video editor.

THINGS I BUILD

Proof of
work.

Real products, experiments, and engineering systems — from voice AI and RAG to full-stack applications.

Featured · FEATURED · VOICE AI

Golden Minutes

A multilingual, interruptible AI voice assistant for cyber-fraud complaint intake that captures incident details, confirms information, generates a PDF report, and enables admin-side review.

THE PROBLEM — Cyber-fraud victims need to report critical incident details quickly and accurately, while valuable response time can be lost during manual information collection.

THE BUILD — Structured conversational intake with Gnani STT/TTS, LLM reasoning, RAG, report generation and admin review.

RESULTS — 96.7% field extraction accuracy · 58/60 evaluation scenarios · 2.1s p95 turn latency.

FastAPIGnani STT/TTSGroq LLMsChromaDBRAGPython
Demo ↗GitHub ↗

SPEECH AI

SpeechTrust

An AI speech-assessment platform that analyzes spoken responses across speech characteristics and semantic relevance, producing a 0–100 assessment score.

40+ modules · 1,000+ questions · 5 speech features.

FastAPIWhisperLibrosaSentence TransformersReactMongoDB
Live Demo ↗GitHub ↗

COMPUTER VISION

VestIQ

A visual fashion discovery system using FashionCLIP embeddings to find visually similar products and support outfit compatibility analysis.

ROC-AUC 0.803 · 5-fold mean ROC-AUC 0.807.

FashionCLIPFastAPINext.jsSupabaseSerpApi
GitHub ↗

REAL-TIME VOICE AI

Voxora

A real-time interruptible voice assistant combining VAD, speech recognition, LLM reasoning, RAG and cancellable TTS for natural turn-taking.

50/50 interruption tests · 0.2ms TTS cancellation · 2.99s average response latency.

Silero VADWhisperGPT-OSS-20BOrpheus TTSRAGFastAPI
Demo ↗GitHub ↗

FULL-STACK

NoteHandle

A full-stack study platform bringing notes, AI assistance, planning, progress tracking, mind maps, checklists and authentication into one workflow.

ReactRedux ToolkitNode.jsExpressMongoDBJWT
Live Demo ↗GitHub ↗

AI + FULL-STACK

Arch AI

An AI-powered architectural planning application that transforms user requirements into floor-plan concepts through an interactive full-stack workflow.

ReactNode.jsExpressMongoDBAI
Live Demo ↗GitHub ↗

FULL-STACK

Wanderlust

A full-stack accommodation platform with property listings, authentication, CRUD workflows, reviews and database-backed application logic.

MongoDBExpressNode.jsEJSREST APIsAuthentication
Live Demo ↗GitHub ↗

MY TOOLKIT

The
stack.

Tools I actually use to move from model and API to interface and shipped product.

Each colour is one of my agents. You’ll find the same colours on the project tags above and on the bots moving around this page.

AI / MLATLAS

Machine LearningGenerative AI & LLMsRAGNLPComputer VisionPyTorchscikit-learnSentence TransformersAudio Feature Extraction

LANGUAGESNOVA

PythonC++JavaScriptTypeScriptSQLJava

BACKENDECHO

FastAPINode.jsExpress.jsREST APIsJWT Auth

FRONTENDPIXEL

React.jsRedux ToolkitTailwind CSS

DATAFLUX

MongoDBPostgreSQLSupabaseAlteryx

TOOLSNEXUS

Git & GitHubPostmanVercelRenderCloudinary

MY EXPERIENCE

Where I’ve
worked.

Oct 2026 – Nov 2026

Machine Learning & AI Intern

BeeSkilled · Remote

  • Built preprocessing pipelines on 800+ records using Pandas and scikit-learn, applying median imputation, LabelEncoder, and one-hot encoding to avoid data leakage.
  • Trained and evaluated Linear Regression, Logistic Regression, and Decision Tree models across California Housing, Titanic, and Heart Disease datasets, achieving 85%+ classification accuracy.
  • Applied K-Means clustering and PCA (4D → 2D) on Iris and saved deployment-ready pipelines with Joblib.
Jun 2026 – Jul 2026

Data Analysis Trainee

Medi-Caps University · Indore

  • Built Alteryx workflows for cleaning, transforming, joining, and preparing datasets across sales, customer, finance, and education domains.
  • Performed data analysis using summaries and averages and prepared outputs for reporting.
  • Applied Alteryx Designer Core and Advanced techniques through hands-on data analysis exercises.

MILESTONES

Progress &
recognition.

Verify
Verify
Verify
Social Internship Programme ↗

Taught AI concepts to students at a government school

Verify
Verify
LeetCode ↗

300+ problems solved

Verify