Backward pass, in miniature

My framework NeuralFabric computes gradients with reverse-mode autodiff. Here is the same idea on one neuron, out = tanh(w·x + b). Move the sliders and watch each gradient (blue) flow back from the output.

1.5
0.8
-0.5

Projects

NeuralFabric

Python, NumPy
Source Code
  • Built a modular machine learning framework from first principles using Python and NumPy, implementing tensor operations, reverse-mode automatic differentiation, and foundational regression models.
  • Developed numerical-gradient tests with Pytest to validate autograd correctness, enforced static typing with MyPy, and structured the project for open-source distribution with modern Python packaging and GitHub Actions.

DoorDarshan

FastAPI, WebSocket, WebRTC
Source Code
  • Engineered a real-time video conferencing platform using FastAPI, WebSockets, and browser-native WebRTC for live audio/video communication, screen sharing, and peer signaling.
  • Implemented JWT-authenticated meeting rooms, participant management, in-meeting chat, and meeting scheduling with PostgreSQL persistence using SQLAlchemy and Alembic.

Kryptika

ECDSA, P2P, HTTP
Source Code
  • Implemented a blockchain from scratch in pure Python, covering SHA-256 hashing, Proof of Work, ECDSA-signed transactions, transaction validation, and longest-chain consensus.
  • Engineered P2P networking, persistent mempool and blockchain storage with SQLite, thread-safe HTTP communication, and a browser-based dashboard with a visual chain explorer.

Timeless

PyTorch, Transformers
Source Code
  • Built a decoder-only Transformer language model from scratch in PyTorch, implementing tokenization, causal self-attention, multi-head attention, residual connections, and normalization.
  • Developed an end-to-end LLM workflow covering pretraining on unlabeled text, model checkpointing, text classification fine-tuning, and instruction tuning for downstream tasks.

AdityaL1-MAG

Python, Signal Processing
Source Code
  • Developed a data-analysis pipeline for ISRO’s Aditya-L1 MAG observations, including data ingestion, filtering, and visualization of spacecraft magnetic-field measurements.
  • Performed magnetic-field characterization and exploratory analysis to identify patterns and variations in the observed environment.

Education

B.Tech, Artificial Intelligence and Data Science, Jabalpur Engineering College

Second Year (Sophomore)

Aug 2025 – Jun 2029
Higher Secondary, PCM + Computer Science (CS), Stemfield International School Apr 2011 – Mar 2025

Certifications

Machine Learning Specialization, DeepLearning.AI, Stanford University

Completed coursework covering supervised learning, neural networks, model optimization and regularization.

Jul 2026
Python for Data Science (Elite + Silver, Top 2% Topper), NPTEL, IIT Madras

Earned the Elite + Silver certification with an 85% score, ranking among the top 2% of candidates nationwide in NPTEL’s Python for Data Science course offered by IIT Madras.

Oct 2026

Awards and Honours

Finalist, RAGNAROK Hackathon (Google Developer Groups)

Selected as a finalist in a competitive hackathon for technical problem-solving and rapid project development.

Jan 2026

Technical Skills

Programming Languages
Python, C, C++, Java, Go, JavaScript, TypeScript, R
Backend Engineering
FastAPI, REST APIs, WebSockets, WebRTC, SQLAlchemy
Databases & Caching
SQL, PostgreSQL, MySQL, MongoDB, Redis
Machine Learning & Data Science
PyTorch, TensorFlow, Scikit-learn, NumPy, Pandas
Cloud & DevOps
AWS, Docker, Kubernetes, GitHub Actions
Systems & Tools
Linux, Bash, Git, GitHub