Hello! Thanks for stopping by.
Explore this creative space and find ideas for your business with us. Crafted by Mere Dadoji
Shashankk Shekar Chaturvedi
×
Download Resume
❅
❅
❆
❄
❅
❆
❄
❅
❆
❄
×
About Me & Experience
Hi there, I’m Shashankk—thanks for visiting my portfolio.
I’m currently pursuing a Master’s degree in Applied Artificial Intelligence at
Stevens Institute of Technology, specializing in advanced machine learning
and data engineering. Before that, I earned my M.S. in Computer Applications
from Amity University, focusing on machine learning, natural language processing,
and intelligent systems.
I worked at UBS as an engineer, contributing to the development and quality
assurance of global internet and intranet applications in agile teams.
Prior to that, I worked with Accenture, where I got hands-on experience
designing and delivering technical solutions across technology, operations,
and consulting projects, all aligned with business needs.
Recently, I had the opportunity to work on the NASA L’Space program, where
I contributed to a hybrid insulation system for extreme space environments,
using materials engineering and simulation tools to support robotic systems
and thermal management. Previously, I've also interned at Spartificial as a Machine Learning
Research Intern, focusing on geospatial analysis and classification mapping.
Currently, I’m also serving as a Student Ambassador for FlutterFlow, exploring and sharing
knowledge about no-code/low-code platforms that make app development more
accessible to everyone. Along the way, I’ve also dabbled in AI+Marketing
projects, integrating real-time insights and behavioral data analysis
to improve user engagement and deliver data-driven strategies.
I’m especially interested in aligning AI-driven solutions with user-centric
design and product management principles. My goal is to bridge technical
innovation and business strategy to create scalable, impactful solutions
that address real-world problems effectively.
Please choose an option:
Recent Projects
Software Engineering/Generative AI
Image credit: In-House
CyberLearn - Cybersecurity Learning Hub
CyberLearn is an interactive, single-page web application built to help users master
essential cybersecurity concepts. It features a landing page with a “Matrix-style”
animated background, along with sections for Vocabulary, Quizzes, Security Scenarios,
and a hands-on Linux Lab—all seamlessly integrated via JavaScript. This hub leverages
Bootstrap, Papa Parse, Animate.css, Google Fonts, and other modern tools to deliver
a dynamic learning experience.
CyberLearn is a full-featured, single-page web application showcasing
crucial cybersecurity topics via an interactive UI. It is split into multiple sections
(Vocabulary, Quiz, Security Scenarios, and a Linux Lab) and uses a JavaScript-driven
approach to switch between sections without reloading the page.
The landing page features a Matrix-rain animation for a striking “hacker” style,
while the main “Learning Hub” covers essential cybersecurity concepts, letting users
test their knowledge and engage with hands-on labs.
Key highlights: a dynamic Vocabulary system loaded from a CSV with Papa Parse,
an adaptive Quiz module, scenario-based problem solving, a fully simulated Linux Lab environment,
and a layered security design (disabling right-click, blocking dev tools usage, etc.) for an
immersive “secure environment” feel.
Core Technologies & Tools:
Bootstrap 5 – for responsive design, layout grids, and modals.
Papa Parse – automatically loads and parses the CSV of cybersecurity terms.
Animate.css – provides smooth CSS animations on the landing page and hero sections.
Google Fonts – “Space Grotesk,” “JetBrains Mono,” and “Noto Sans Devanagari” for a
modern hacker-themed aesthetic.
OpenAI API (planned) – included as a script reference for potential future AI-driven features.
Vanilla JavaScript – handles quiz logic, section toggling, scenario interactions, and
a “fake” file system for the Linux lab.
Feature Breakdown
CyberLearn is built around several distinct modules that together deliver a comprehensive
cybersecurity learning experience.
1. Landing Page & Matrix Rain
The landing page uses a canvas-based “Matrix rain” animation (inspired by the
classic “falling green text” effect). JavaScript renders random characters,
including Sanskrit, binary, hex, and symbols, giving the site a futuristic hacker vibe.
A loading screen with a progress bar simulates system initialization, and security
blocks (e.g. no right-click, F12) increase the sense of a “secure environment.”
2. Vocabulary Section
The vocabulary is loaded from a CSV (“cybersecurity-vocabulary.csv”) using Papa Parse.
Each term is displayed in a Bootstrap card with a short definition; clicking it opens
a modal showing expanded info (synonyms, sources, extended definitions, etc.). A search
bar filters the terms in real time.
3. Interactive Quiz
Three quiz modes (Term→Definition, Definition→Term, and True/False) let users test
their knowledge with randomly selected questions from the CSV data. A progress bar
displays how far along the user is, with final results showing correct answers vs. total.
4. Security Scenarios
A mini “scenario game” presents short cybersecurity challenges (like phishing emails
or public Wi-Fi usage). Users choose from multiple responses, immediately seeing if
they are correct (highlighted in green) or incorrect (red).
5. Linux Lab
A simulated “terminal” environment teaches Linux commands. Users move through
progressive “chapters,” each requiring specific commands (like pwd,
ls, cd, rm, etc.). Once the correct
command is entered, the user unlocks the next task or chapter. This portion
uses a custom, in-memory “fake file system” object to mimic basic shell behaviors.
Technical Roles & Functions
Module/Component
Purpose
Matrix Rain Canvas
Renders an animated “hacker” aesthetic background using custom JavaScript to draw random characters in columns.
Vocabulary Loader
Uses Papa Parse to fetch CSV terms, then dynamically generates Bootstrap cards and modals for user exploration.
Quiz Engine
Creates dynamic quizzes based on the loaded vocabulary, supports various question formats, and tracks user score.
Scenarios “AI Game”
Presents short security challenges with correct/incorrect responses, reinforcing best practices.
Linux Lab Simulation
Mimics a terminal environment where users issue commands to complete tasks, gradually unlocking advanced chapters.
Why This Matters
Hands-On Cybersecurity Education: Integrates reading, quizzes,
and terminal practice to deepen understanding.
Immersive Design & Aesthetics: The Matrix-themed background and
security overlays foster an engaging “hacker” atmosphere.
Modular & Extendable: The code is structured as separate sections,
making it easy to add new lessons, labs, or advanced features (like AI).
Data-Driven Content: CSV-based vocab and scenario definitions let
non-technical contributors expand the repository of terms and scenarios.
Image credit: In-House
Task∞Wire AI - Intelligent Scheduling Assistant
Task∞Wire AI is an innovative task management application that leverages advanced AI to optimize your daily schedule.
Built with React, TypeScript, and Vite, it learns your habits and energy levels to suggest optimal time slots and automatically
detects scheduling conflicts—reducing them by 85%—all while delivering a smooth, responsive user experience.
TaskWire AI is a cutting-edge scheduling assistant designed to streamline task management by leveraging AI and modern web technologies.
The application intelligently analyzes user behavior, habits, and energy levels to suggest the best time slots for tasks, while its real-time conflict detection
and resolution mechanism ensures a smooth daily workflow. This comprehensive solution integrates a robust tech stack including React, TypeScript, Vite, Tailwind CSS,
and advanced AI capabilities powered by OpenAI.
Key Features of TaskWire AI include:
AI-Powered Scheduling
Utilizes a custom AI algorithm with OpenAI integration to analyze user patterns and energy levels, reducing scheduling conflicts by 85%.
Natural Language Processing
Enables intuitive task creation via plain language commands (e.g., "Meeting at 3 PM every Monday") for a seamless user experience.
Responsive & Performant UI
Built with React and Tailwind CSS, it achieves a 95% Lighthouse performance score and a 60% reduction in bundle size through efficient code splitting and lazy loading.
Modular Architecture
Designed with reusable components and custom hooks, ensuring maintainability and scalability, backed by 98% test coverage using Vitest.
Progressive Web App (PWA)
Offers offline capabilities and dynamic state management for an uninterrupted user experience.
Benefits & Impact:
Boosts productivity by intelligently reducing scheduling conflicts and optimizing daily workflows.
Enhances user experience with a responsive, intuitive interface and natural language commands.
Ensures a scalable and maintainable solution capable of handling increasing task volumes efficiently.
Technical Breakdown
TaskWire AI integrates a suite of modern technologies to deliver a seamless scheduling experience. Its architecture is divided into distinct layers, each optimized for performance, scalability, and user interaction.
Key Components of the Architecture
1. Frontend
React, TypeScript & Vite: Provides a robust and scalable foundation for building dynamic user interfaces. Tailwind CSS ensures a sleek, responsive design with efficient code splitting and lazy loading.
2. AI & Scheduling Engine
Custom AI Algorithm & OpenAI Integration: Analyzes user behavior, energy levels, and task patterns to generate optimal time slot suggestions, powering real-time conflict detection and resolution.
3. Natural Language Processing (NLP)
NLP Interface: Allows users to effortlessly create and manage tasks using plain language, streamlining the scheduling process.
4. Modular Architecture & Testing
Reusable Components & Custom Hooks: Ensures a maintainable and scalable codebase, reinforced by 98% test coverage using Vitest.
5. Progressive Web App (PWA) Capabilities
Offline & Dynamic State Management: Provides uninterrupted task management through proactive caching and efficient state handling.
Technical Roles
Component
Functionality
AI Scheduling Engine
Analyzes user patterns and energy levels to suggest optimal time slots, reducing scheduling conflicts by 85%.
Real-Time Conflict Detector
Monitors scheduling overlaps and provides immediate, AI-suggested alternative times.
Responsive Frontend
Delivers a smooth, interactive user experience using React, TypeScript, and Tailwind CSS.
Modular Architecture
Ensures a scalable, maintainable solution through reusable components and custom hooks, with robust testing via Vitest.
PWA Capabilities
Offers offline access and dynamic state management for continuous task scheduling.
Why This Architecture Matters
Enhanced Productivity: Intelligent scheduling minimizes conflicts and streamlines workflows.
User-Centric Design: An intuitive NLP interface and responsive UI ensure ease of use for all users.
Scalability & Maintainability: A modular, well-tested architecture can grow with your needs.
Future-Proofing: Advanced AI integration and PWA capabilities position TaskWire AI for ongoing innovation.
Data Engineering (End-to-End Project)
Image credit: In-House
PM2.5 Real-Time Dashboard & Pipeline
This solution continuously monitors PM2.5 air pollution in real time, flags anomalies, and archives
historical snapshots for deeper trend analysis. A 2-minute rolling buffer keeps memory usage low in
“Live” mode, while aggregated snapshots (average/max PM2.5, hazard/normal counts) are stored in S3
for “Historic” analysis. IsolationForest detects out-of-range pollution levels (“HAZARD”), and a
GPT-based Q&A interface makes data accessible to everyone.
This project tackles the critical challenge of air quality monitoring by providing real-time PM2.5
pollution analytics and in-depth historical trend analysis, designed to aid researchers,
policymakers, and the general public. The architecture integrates streaming data pipelines,
anomaly detection, forecasting models, and natural language processing into a single, scalable
system, enabling real-time decision-making and long-term insights.
Key features of the system include:
Real-Time Analytics with Minimal Memory Footprint
Utilizes a 2-minute rolling buffer to deliver real-time PM2.5 monitoring without overloading memory,
ensuring the system is lightweight and efficient.
Historical Aggregator Insights (S3-Based Snapshots)
Stores periodic aggregated metrics (average PM2.5, maximum PM2.5, hazard/normal counts) in AWS S3
for in-depth historical trend analysis, enabling researchers to analyze pollution patterns over time.
Advanced Anomaly Detection
Implements IsolationForest to identify anomalies and classify them as “HAZARD,” helping pinpoint
abnormal pollution levels and their geographic distribution.
Synthetic Sensor Data Generation
Simulates real-world PM2.5 sensor data streams to test and validate the system under various
conditions, ensuring robustness in real-world deployments.
Interactive GPT-Based Q&A
Leverages GPT-3.5 Turbo to answer natural language queries about air quality trends, hazard levels,
and region-specific insights, making complex data accessible to non-technical users.
Region-Based Mapping & Forecasting
Visualizes PM2.5 levels on an interactive map, with region-based aggregation for better geographical
context, and time-series forecasting for proactive environmental measures.
Research-Oriented Analytics
Integrates advanced computations in historical data analysis (regional pollution distribution,
hazard trends, forecasted air quality), empowering researchers with granular insights into
pollution sources and impacts.
Impact & Applications:
Environmental Monitoring: Aids environmental agencies in tracking pollution hotspots in real
time and assessing long-term patterns.
Policy Decisions: Provides actionable insights for urban planning, pollution control, and
public health advisories.
Educational Tool: Offers a learning platform for students and researchers exploring data
engineering, environmental science, and machine learning.
Technical Breakdown
This project also integrates real-time analytics, anomaly detection, and
machine learning for PM2.5 air quality monitoring. It provides both
immediate insights (2-min buffer) and long-term data analysis (S3 aggregator),
ensuring scalability and advanced forecasting capabilities.
Key Components of the Architecture
1. Data Ingestion
Kafka Producer: Sends synthetic PM2.5 sensor data (PM2.5, timestamp, lat/lon, sensor ID)
into the air_quality Kafka topic in real time.
2. Aggregation Layer
Aggregator Service: Consumes messages from air_quality, computes average/
max PM2.5, hazard vs. normal counts over 3–5 minute intervals, then stores JSON snapshots in S3
for historical trend analysis.
3. Anomaly Detection
IsolationForest Model: Flags outliers as “HAZARD” if readings deviate significantly
from typical ranges, aiding immediate alerts and mapping.
4. Forecasting
Predictive Analytics: Employs linear/trend-based or advanced ML to forecast
hazard vs. normal counts and pollution patterns.
5. Visualization & User Interaction
Dash App: Provides “Live” (2-min rolling) or “Historic” (aggregator-based) data
views, region-based charts, hazard tagging, and GPT-based Q&A for user-friendly queries.
6. Containerization
Dockerized Pipeline: Kafka, aggregator, and Dash app run in separate containers,
ensuring scalable, consistent deployments.
Technical Roles
Concept
Functionality
Real-Time Streams
Kafka fosters near-zero-latency ingestion from synthetic sensors to aggregator consumers.
Aggregator Snapshots
Summarized data (avg, max, hazard) in S3 keeps storage minimal yet historically insightful.
Anomaly Detection
IsolationForest categorizes abrupt pollution spikes as “HAZARD” for immediate attention.
Forecasting
Predicts near-future air quality or hazard trends, assisting proactive environmental measures.
GPT Q&A
Allows user-friendly queries on hazards, historical stats, or region-based distribution,
bridging data science and layperson accessibility.
Why This Architecture Matters
Scalable & Cost-Efficient: Minimal aggregator snapshots keep memory usage
and S3 costs low, while containerization scales effortlessly.
Real-Time + Historic Insights: Continuous sensor ingestion plus S3-based
historical queries reveal comprehensive trends over time.
Policy & Research Impact: Hazard mapping, forecasting, and anomaly detection
guide environmental strategies and health advisories.
Accessible Data: GPT Q&A fosters a natural-language interface, lowering the
barrier for wide audiences to interpret complex data.
Machine Learning / Modeling and Optimization
HybridLex: Advanced Plagiarism Detection System
Developed a sophisticated plagiarism detection tool that combines lexical fingerprinting with transformer-based semantic embeddings to accurately identify and quantify content duplication. Libraries used:- streamlit, nltk, sentence-transformers, scikit-learn, numpy, pandas, torch, transformers
Revert Smart AI – Email Follow-Up Tracker (Testing)
Developing a cross-platform productivity app using FlutterFlow, integrating Firebase for authentication and real-time reminders, GPT API for AI-powered email drafts, and Lottie animations for an interactive onboarding experience.
The NASA L’Space program is a hands-on initiative designed to prepare students for real-world space
challenges. Last Fall in 2024, I had the opportunity to participate in this program, where I received
training and mentorship from NASA scientists and engineers. My team focused on developing a research
proposal that addresses the extreme temperature swings and harsh conditions found on other planetary
bodies, such as Mars and Venus. Our goal was to explore a robust insulation and habitat design—one that
not only benefits spacecraft but also supports potential in-situ habitats, ensuring a stable,
life-sustaining environment for future missions and crew.
In my role as a simulations and modeling engineer, I conducted thermal simulation studies to validate
our proposal’s feasibility. This involved evaluating advanced insulating materials, modeling
environmental stress factors, and refining designs for an integrated habitat concept. By leveraging
computational analyses, we aimed to create a more efficient hybrid insulation system capable of
shielding crucial mission components from the severe conditions found on planetary surfaces. Overall,
the NASA L’Space experience sharpened my skills in research proposal development and advanced modeling
techniques, while giving me deeper insight into the engineering innovations that help shape human
exploration of distant worlds.
Certificate link: Coming Soon
Stellar Classification - A Particle Swarm Optimization Approach
This project employs Particle Swarm Optimization (PSO) to automate and optimize the spectral classification of celestial bodies, enhancing our understanding of the universe's composition and evolution.
Credit Card Fraud Detection Using Machine Learning Algorithms
Our project explores optimal Machine Learning algorithms like Logistic Regression, Decision Trees, and XGBoost to tackle credit card fraud in online transactions. Through techniques like Random Oversampling and SMOTE, we aim to enhance ML-based solutions for more secure online transactions.
This project involves developing a stock market prediction tool utilizing machine learning algorithms such as ARIMA, LSTM, and Linear Regression to forecast stock prices and analyze sentiment from tweets.
NutriScan is a mobile app designed to help users make informed decisions about their health and dietary needs while shopping. Users can register, input health details, and scan product labels to check suitability.
Developed a comprehensive marketing strategy for Neuradent, a smart toothbrush brand. The project focused on leveraging AI and data-driven insights for personalized oral care, real-time feedback, and effective marketing channels.
Key components: Data-driven oral care, personalized coaching, integration with the NeuraDent app, AI in marketing and product development.
IoT-Based GPS Tracking System Using Raspberry Pi 4B
Overview: Developed an advanced IoT-based GPS tracking system for
real-time vehicle or personal movement monitoring. This setup harnesses a
Raspberry Pi 4B, ensuring high performance while maintaining portability.
Hardware Components:
• Raspberry Pi 4B: Core microcontroller with Raspbian OS.
• GPS Module: High-sensitivity for accurate location tracking.
• Portable Wi-Fi Dongle: Real-time data transmission to IoT dashboards.
Project Details:
System Design & Setup: Configured Raspberry Pi 4B with the latest Raspbian OS,
creating a static IP for remote access. Automated GPSD on boot using custom Linux scripts.
Programming & Data Processing: Wrote Python scripts to collect and process GPS
data in real time. Integrated with Ubidots IoT dashboard to visualize and monitor location
info on web and mobile.
Data Visualization & IoT Integration: Built a custom dashboard for tracking vehicle
coordinates on Google Maps, plus real-time alerts and data analysis tools to enhance accuracy
and user experience.
Learning Outcomes:
Gained hands-on experience with embedded systems, IoT integration, and real-time data
visualization. Strengthened skills in Python programming, GPS technology, and designing robust
hardware–software systems from the ground up.
Automated Network Traffic Analysis and Monitoring System
Developed a system for automated network traffic analysis to detect anomalies and ensure optimal performance. The project involved capturing network traffic data, performing real-time analysis, and visualizing the results on a monitoring dashboard. This system enhances network security and performance by identifying potential threats and inefficiencies.
This is a protytpe design of a Travel app designed to give the glimpse of the actual app. This App was designed using Adobe XD.
Harmony - The Music App
This is the prototype design of Harmony, which is a music app. This design gives a good glance over the App UI and was made using Adobe XD.
Poster Designing
Nature Workshop
Want to get amazing Posters like this for your Business? Click the following link for more details:- More Details
Hour of Code
Want to get amazing Posters like this for your Business? Click the following link for more details:- More Details
Certifications
Goldman Sachs Program
Participated in a virtual program focused on finance‐related
software solutions and real‐world technical challenges. Gained exposure to real-world finance and software solutions, strengthening algorithmic thinking and problem-solving skills.
McKinsey Forward Program
Sept 2024 – Dec 2024
Learned structured problem-solving, effective communication, and leadership skills to tackle dynamic business challenges using advanced analytical frameworks.
NASA Open Science
Demonstrated a deep understanding of open science principles, integrating publicly accessible datasets and transparent software development into real-world NASA research projects.
Oracle Database 11g
Attended workshops covering DBA tasks, database security, backups, and performance tuning for enterprise-scale data, developing hands-on expertise in managing data environments.
Walmart USA Advanced S.E.
Built foundational software engineering skills including designing data structures, UML diagrams, and system models, aligning code efficiency with large-scale commercial needs.
IBM Applied AI Developer
Built and deployed AI chatbots, image classifiers, and real-time analytics workflows using Watson APIs. Strengthened Python development and integrated modern AI services to solve business challenges.
Machine Learning (Stanford)
Mastered core ML algorithms through a combination of theory and practice, gaining insights into model evaluation and optimization techniques for diverse data science applications.
