BS in Applied AI & Data Science student @ IIT Jodhpur
Building robust, intelligent applications at the intersection of deep neural analytics, mathematical modeling, and production-grade software infrastructures.
My engineering foundation is built on absolute logical rigour. Forged in the intense residential environments of JNV and Sikar, I approach data science from a mathematical, first-principles perspective.
Currently pursuing a BS in Applied AI & Data Science at IIT Jodhpur, I bridge the gap between abstract mathematics and production-grade software. I specialise in machine learning, full-stack web development, and Android architectures. To me, a model is only as powerful as the infrastructure that serves it. I don't just train architectures—I build the real-world pipelines that make them scale.
Consistent high performance in rigorous technical curriculum.
Transformer architectures and large language models.
I don't just train models in Jupyter notebooks. I containerize, deploy, and monitor them using industry-standard MLOps practices.
Currently pursuing a BS in Applied AI & Data Science with a strong focus on Machine Learning, Artificial Intelligence, Statistics, Programming, and Data Analytics.
Chasing a better version of tomorrow
ML • AI • Android • Web
A systematic progression through rigorous academic environments, building the mathematical and algorithmic foundation required for advanced AI research.
Focusing entirely on deep learning architectures, scalable data pipelines, automated machine learning (MLOps), and advanced statistical algorithms.
JEE Preparation and 12th board diagnostics execution. Intensive quantitative problem solving.
11th standard systematic progression under residential system criteria.
6th - 10th standard foundational reasoning modules execution.
Advanced algorithmic optimization, data structuring, and backend scripting.
Architecting CNNs, RNNs, and custom neural networks for predictive analysis.
Data parsing, feature engineering, and high-dimensional statistical modeling.
Integrating Large Language Models, semantic parsing, and RAG architectures.
Current CGPA
Projects
Technologies
Started AI Journey
A selection of end-to-end applications demonstrating my ability to build robust machine learning models, from scratch frameworks, to applied GenAI tools.
Most ML coursework treats backprop as a library call. I wanted to know what model.fit() actually does under the hood — so I built the engine myself: a PyTorch-style Sequential container with Dense layers, ReLU, Softmax + Cross-Entropy, and SGD/Adam optimizers, all hand-derived from matrix calculus in raw NumPy. Trained end-to-end on MNIST, it reaches ~97% test accuracy — with zero PyTorch or TensorFlow anywhere in the stack.
Feeding a whole legacy file to an LLM in one shot reliably produced hallucinated variables and truncated logic — context windows just aren't built for large-scale refactors. My fix: a Python backend that parses each file into an Abstract Syntax Tree first, breaks it into self-contained semantic chunks, translates each in isolation, then re-stitches the result — preserving type information and structure throughout. Ships as a real VS Code extension.
Manual PR review doesn't scale — reviewers skim, security issues slip through, and feedback loops stretch for days. Aegis AI turns a pull request into a structured engineering audit instead: an overall score, risk level, severity-classified findings (with confidence ratings, not just "this looks wrong"), and fixes linked directly to the issue they address. Gemini + LangGraph drive the analysis; a FastAPI/Celery/Redis pipeline handles webhooks asynchronously so GitHub never times out waiting on the AI, and a Streamlit dashboard surfaces review history and trends from Supabase.
AI-powered DevSecOps code review platform that scores GitHub pull requests and standalone code for security, performance, and quality using Gemini and LangGraph.
Android-based artificial intelligence application for scanning, parsing, and scoring applicant resumes automatically.
A recommendation engine designed to intelligently match students with optimal curriculum paths based on historical data.
Conversational AI bot configured to dynamically interview candidates and extract relevant competency metrics.
Integration and deployment of Large Language Models (LLMs) via API to a scalable edge network on Vercel.
IIT Jodhpur
Participating in an intensive 3D printing and additive manufacturing course at IIT Ropar alongside classmates, focusing on hands-on hardware prototyping.
Proud to be a Navodayan. Sharing this memorable Class 10 group photo with the peers and mentors at JNV Fazilka who fostered our growth, discipline, and problem-solving skills.
H
M
Stage.
Swipe to explore
Active open-source contributions, research logs, and version control mastery. My GitHub is the public ledger of my daily coding, data engineering, and deployment habits.
Engaging deeply with IIT Jodhpur curriculum data science projects, focusing on statistical modeling optimizations.
Documenting machine learning roadblocks, architectural decisions, and MLOps strategies for future reference.
Comprehensive mastery of core supervised and unsupervised algorithms.
Verified CredentialAdvanced data analysis, cleaning methodologies, and structuring pipelines.
Verified CredentialProven ability to rapid-prototype, build, and deploy ML models under strict deadlines.
Competition LogGoogle AI Studio
Google Cloud Skills Boost
Kaggle
IIT Jodhpur Coursework
Upcoming Certifications
I am actively seeking machine learning internships, open-source collaborations, and rigorous data science roles.
Whether you are building scalable AI architectures, need an Android integration, or just want to discuss transformer models—my inbox is always open. Let's engineer something impactful.