System Status: Online & Building

Jashan Tiwari

Passionate about AI, full-stack development, and innovation

Explore Architecture GitHub

System Overview

Biography

Operating out of Patiala, Punjab, I’m a part-time football player and full-time coder passionate about building intelligent, scalable, and impactful technology solutions. As a motivated and detail-oriented technical architect, I bridge the gap between advanced machine learning and modern full-stack development to solve real-world problems. Currently pursuing the BS in Data Science & Applications program at IIT Madras, I specialize in optimized ML pipelines, secure backend systems, scalable APIs, and AI-driven applications. I have developed projects including multi-user platforms, AI-powered systems, and data visualization applications using real-world APIs. Passionate about innovation, artificial intelligence, and scalable software engineering, I continuously explore emerging technologies to build secure, efficient, and future-ready digital solutions.

Databanks

Intern

Cystar, IIT Madras

Present

BS in Data Science

IIT Madras

Expected 2027

Diploma Data Science

IITM

Completed

Diploma Programming

IITM

Completed

AI & Data Science

Python
Scikit-learn
XGBoost
LightGBM
Transformers
GANs

Backend & Cloud

Django
FastAPI
Flask
Node.js
PostgreSQL
Redis
Celery

Mobile & Frontend

Kotlin (Android)
Vue.js 3
JavaScript
Bootstrap 5
Chart.js

Project Architecture

Cybersecurity / Mobile SDK

FraudGuard SDK

A sophisticated Android SDK for real-time fraud scoring and agentic firewalling, securing orchestration and monitoring behavioral signals.

  • Implemented root and emulator detection architectures.
  • Behavioral signal monitoring including sensor data and typing cadence.
  • Integration with NeMo Guardrails for enhanced LLM security.
  • Multi-agent orchestration utilizing Scion.
Kotlin
Android SDK
AI Firewalls
Machine Learning / NLP

Predictive Models & Sentiment Analysis

High-scoring competitive modeling built for Kaggle challenges, focusing on deep learning architectures and rigorous hyperparameter tuning.

  • Comment Category Prediction Challenge architecture.
  • Advanced Movie Sentiment Analysis utilizing optimized deep learning.
  • Model optimization utilizing Adam/AdaGrad optimizers.
XGBoost
LightGBM
Transformers
Scikit-learn
Full-Stack Application

Quiz Master V2

A multi-user exam preparation platform with role-based access control, quiz management, and real-time analytics dashboards.

  • RBAC with JWT authentication handling Admins & Users.
  • Advanced Analytics rendering with Vue.js & Chart.js.
  • Background Jobs via Celery & Redis for scheduled reporting.
Flask
Vue.js 3
Redis
Celery
Data Analytics

Business Performance Engine

Quarterly performance and revenue analysis system developed for a commercial grocery enterprise (Guru Ji Da Modikhana, Punjab).

  • Comprehensive revenue tracking and quarterly forecasting.
  • Data visualization and professional reporting generation.
  • Strategic planning insights derived from historical store data.
Data Analysis
Business Intelligence
Web Application

Household Services Platform

A multi-role service platform orchestrating admins, service professionals, and customers for home servicing solutions.

  • Dynamic rendering of localized pages using Jinja2 templates.
  • Secure credential and request management via Werkzeug.
  • Robust backend processing and job allocation logic.
Flask
SQLAlchemy
Jinja2