kanhaiya sharma

+•• ••••• •••••••••••@•••••.com

Neural Digital Twin for Early Detection of Cognitive Disorders · Independent Researcher · 1.8 yrs

Ask about kanhaiya

Proof of Experience

Neural Digital Twin for Early Detection of Cognitive Disorders

Independent Researcher · Aug 2025 – Present

  • Developed multimodal neural digital twin integrating EEG (TUH) and MRI (ADNI) for early cognitive decline detection and tumor classification.
  • CMF-VIT EEG model achieving 88-92% accuracy with 70-90% class sensitivity on 22 patient records.
  • Implemented EfficientNet-B4 MRI classifier with 97–98% validation accuracy and 97.6% test accuracy on 111 records.
  • Designed spatio-temporal CNN/LSTM pipelines with automated evaluation and visualization for progression analysis.

Legal Document Analysis Tool

Independent Developer · Jun 2025 – Present

  • Built secure legal document platform using Next.js, Supabase (pgvector), and Auth0 for managing 1,000+ documents.
  • Accomplished ~50ms upload/retrieval with signed URLs and enforced JWT-based user-level security.
  • Integrated LLM APIs for clause extraction, compliance verification, and precedent search (<1s response time).
  • Deployed edge APIs with SWR-driven frontend for real-time analysis.

Team Task Manager | Full-Stack Web Application

Independent Developer · May 2026 – May 2026

  • Developed a full-stack task management platform using Next.js, Node.js, Express.js, and MongoDB Atlas with role-based Admin and Member access.
  • Implemented JWT authentication, role-based access control, and 15+ REST APIs for project and task workflows.
  • Responsive dashboard UI with real-time task tracking and optimized API responses to under 200ms for core operations.
  • Deployed the application on Vercel and Railway with secure cloud database integration and production-ready API connectivity.

LSTM Stock Market Analysis

Independent Researcher · Jul 2024 – Nov 2024

  • Engineered 3-layer LSTM model with dropout, batch normalization, and L2 regularization.
  • Achieved R2 = 0.96, RMSE = 66.5, MAE = 31.06 on 5,000+ data points.
  • Reduced validation loss by 70% using EarlyStopping and learning rate scheduling.
  • Improved forecasting stability with 30-day lookback and MinMax scaling.
  • Visualized outputs with Matplotlib to track trends, enabling a 10% decrease in forecast deviation and improved interpretability.

Proof-of-Work

No completed sessions yet. Anything kanhaiya finishes shows up here with the evidence behind it.

Proof of Education

B.Tech, Computer Science and Engineering

Kalinga Institute of Industrial Technology · 2022 — 2026

High School Diploma, General Studies

Heritage Academy High School · 2019 — 2021

Secondary School Certificate, General Studies

Heritage Academy High School · 2017 — 2019

A Proof CV — one profile, kept current, with the proof attached. Back to the portal