Building intelligent systems that turn ideas into real-world applications.

I'm a Computer Science & Engineering (Data Science) student focused on building production-oriented AI/ML systems, Generative AI applications, and scalable software.

• 🤖 Building with Machine Learning, Deep Learning & Generative AI
• 🧠 Working with LLMs, RAG, Transformers, Embeddings & Vector Databases
• 🔬 Researching Deep Learning and Time-Series Modeling
• ⚙️ Building end-to-end ML pipelines and AI-powered applications
• 🌐 Interested in Software Engineering, Backend Development & System Design
• 🌱 Currently exploring Agentic AI, MCP, AI Agents & advanced AI architectures
• 💻 Regularly solving Data Structures & Algorithms
• 🌍 Active Open Source Contributor
• 🏆 Hackathon finalist and award recipient
• 📚 Research paper accepted for publication in a Scopus-indexed journal

My goal: Build intelligent, reliable and scalable systems that solve meaningful real-world problems.

Concepts: Supervised Learning • Unsupervised Learning • Classification • Regression • Ensemble Methods • Feature Engineering • Model Evaluation • CNN • RNN • LSTM • GRU • BiLSTM • Attention • Transformers

Focus: LLMs • Prompt Engineering • RAG • Embeddings • Vector Search • Vector Databases • Fine-Tuning • AI Agents • Agentic AI • MCP

Top 10 Teams

Recognized among the top-performing teams in AMD Hackathon 2026.

Finalist — GTU Innovation Competition

Selected as a finalist in the GTU STEMSpire innovation competition.

Finalist — Top 250 / 19,000+ Teams

Selected among the top 250 teams from more than 19,000 participating teams.

Finalist

Selected as a finalist in India's national-level innovation and hackathon program.

Academic Excellence Award — Awarded Twice

Recognized twice with the Dewang Mehta IT Award for Academic Excellence through GTU.

Scopus-Indexed Journal — Accepted

Research paper accepted for publication in JETIA International Journal:

"Adaptive Gated Ensemble BiLSTM with Attention (GEBiLSTM-Attn) for Aero-Engine RUL Prediction"

GirlScript Summer of Code 2025

• Contributed to 5+ open-source repositories
• Multiple merged pull requests
• Certificate of Appreciation

Winter of Code 2026

• Contributed to Ak-dskit
• Enhanced EDA and feature-engineering utilities
• Worked through pull requests, code reviews and issue resolution

My research focuses on predicting the Remaining Useful Life (RUL) of aircraft engines using deep learning.

A hybrid architecture integrating:

• CNN
• GRU
• BiLSTM
• Gated Feature Enhancement
• Attention Mechanisms

The project includes an end-to-end pipeline covering preprocessing, sequence generation, training, hyperparameter optimization and evaluation.

Deep learning system for Remaining Useful Life prediction using NASA CMAPSS datasets.

Tech: Python • TensorFlow/Keras • CNN • GRU • BiLSTM • Attention

🔗 View Repository

Production-oriented ML monitoring platform designed to detect data distribution shifts in deployed models.

Highlights:

• Monitors 284K+ transactions
• Detects drift across 8 critical features
• Benchmarked 5 ML models
• XGBoost fraud detection model with 0.977 AUC
• FastAPI backend
• Interactive Streamlit dashboard
• Automated retraining alerts

Tech: Python • XGBoost • LightGBM • FastAPI • Streamlit • SciPy

🔗 View Repository

Custom Transformer-based language model designed for next-word prediction.

Tech: Python • Transformers • TensorFlow • NumPy

🔗 View Repository

An intelligent financial assistant using Retrieval-Augmented Generation to analyze financial documents and provide grounded financial insights.

Highlights:

• PDF & CSV ingestion
• Semantic retrieval
• Gemini embeddings
• ChromaDB vector store
• Dual knowledge sources
• Automated financial analytics

Tech: Python • LangChain • Gemini • ChromaDB • RAG • Pandas

🔗 View Repository

A production-ready full-stack placement preparation platform that bridges the gap between juniors and placed seniors through real interview experiences, company-wise preparation, and data-driven insights.

🚀 Live Product: Used by 20+ real students and actively maintained.

Tech: React • Tailwind CSS • Node.js • Express.js • MongoDB • JWT • Recharts

🔗 Live Demo • View Repository

StreamAPI is a scalable RESTful backend for a video-streaming platform with JWT authentication, video management, comments, likes, subscriptions and tweets.
Built with Node.js, Express.js and MongoDB, with Cloudinary for media storage and a complete Postman API collection for testing.

Tech: Node.js · Express.js · MongoDB Atlas · JWT · Bcrypt · Cloudinary · Postman

🔗 View Repository

Building, contributing, learning and shipping — one commit at a time.

Agentic AI
    ↓
AI Agents
    ↓
Tool Calling
    ↓
MCP
    ↓
RAG + Long-Term Memory
    ↓
Production AI Systems

Currently interested in building AI systems that can reason, retrieve information, use tools and execute multi-step workflows.

I regularly practice:

• Data Structures & Algorithms
• Arrays & Strings
• Linked Lists
• Stack & Queue
• Sliding Window
• Two Pointers
• Binary Search
• Hashing
• Recursion
• Trees & Graphs
• Dynamic Programming
• SQL & Database Problems

"Turning ideas into intelligent systems."

"Jack of all trades, master of ML, DL & Generative AI — I believe versatility is a strength, and mastery is a choice."