Aakaash M S

Think Board

This full-stack MERN application allows users to manage notes with a clean, responsive UI. It features a RESTful API with proper error handling, Upstash Redis-based rate limiting for security, and is deployed live on Render.

React.js HTML5 Tailwind CSS Node.js Express.js MongoDB Atlas Upstash RESTful API Swagger Render

Cine Verse

A sleek and responsive React app to browse trending movies, search titles, and explore content using the OMDb API. The app features a user-friendly interface with Tailwind CSS for styling, and utilizes Vite for fast development and build processes. It also integrates Appwrite for data management, ensuring a seamless experience across devices.

React Tailwind CSS JavaScript Node.js Vite Appwrite

Subscription Tracker

A production-ready RESTful API with full CRUD operations for managing users and subscriptions, secured with JWT authentication, rate-limited by Arcjet, and integrated with QStash for scheduling workflow reminders. Added interactive and complete API documentation for developers.

Node.js Express.js MongoDB Atlas REST API Postman Swagger Arcjet Upstash Docker

Multivariate Stock Market Forecasting

Developed an LSTM-based model for multivariate stock price prediction using Google stock data. Integrated SHAP for explainability, Power BI and TensorBoard for visualization, and deployed via FastAPI on AWS with a Streamlit app for real-time forecasts.

Python Pandas Matplotlib Tensorflow Keras LSTM PowerBI Streamlit

Sales Forecasting and Churn Prediction

Build a full-stack data science project that forecasts customer sales and predicts churn using machine learning, visual analytics, and interactive dashboards. Containerized with Docker and production-ready via a FastAPI-based REST API.

Python Pandas Matplotlib Seaborn Scikit-learn KMeans Clustering FastAPI Streamlit Docker

Deep Learning

This project provides a comprehensive overview of deep learning, covering essential concepts such as gradient descent, loss functions, and activation functions. It includes practical implementations of core neural network architectures including Multi-Layer Perceptrons (MLPs), Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Long Short-Term Memory networks (LSTMs).

Python Gradient Descent Loss Activation Functions Multi-layer Perceptron Convolutional Neural Network Recurrent Neural Network Long-Short Term Memory Autoencoder

Fake News Detection Using BERT

Developed a fake news detection system leveraging BERT’s transformer architecture for enhanced context understanding, achieving 90%+ accuracy in distinguishing real and fake news.Engineered a comprehensive NLP pipeline, including text preprocessing, tokenization, and fine-tuning, ensuring optimal model performance and scalability.

Python Pandas Matplotlib BERT Hugging Face

Customer Churn Analytics Dashboard

Designed an interactive Power BI dashboard analyzing 7,000+ customer records to track Churn Rate (27%), Revenue ($456K), and retention KPIs. Used data modeling, DAX, and segmentation to enable real-time insights and identify high-risk churn segments.

PowerBI DAX KPI Slicers

Data Driven A/B Testing

Conducted a Chi-Square hypothesis test to analyze differences in mobile operator preferences between users from Tamil Nadu and other regions. Cleaned and segmented raw user data into control and test groups, and visualized behavioral metrics using Matplotlib and Seaborn. Interpreted results with 95% confidence to derive actionable insights, supporting data-driven decisions to improve student engagement and learning outcomes.

Python Pandas Matplotlib SciPy Chi-Square test A/B Testing

CSV-to-Data Warehouse ETL Pipeline

Built a robust ETL pipeline using Apache Airflow to automate the daily ingestion, transformation, and loading of structured sales data from CSV files into a PostgreSQL data warehouse. The workflow was containerized using Docker Compose, integrating Airflow and PostgreSQL services to simulate a production-like orchestration and storage environment.

Python Pandas Apache Airflow PostgreSQL Docker

Hospital Data Warehouse

This data warehouse project demonstrates SQL and data modeling skills through a star schema structure designed to track and analyze hospital visit records, patient data, diagnoses, and readmission trends. The project includes designing a dimensional model for healthcare data, simulating hospital visit data, and building a foundational data warehouse for advanced analytics. The process follows multiple phases: Schema Definition, Data Ingestion, Exploration, Visualization, ETL (Extract, Transform, Load), and Analysis.

Python Pandas MySQL Matplotlib

MultiCloud StockCast

Built a MultiCloud Stock Forecasting system in collaboration with a team of two, utilizing AWS and Azure services to predict stock prices using LSTM models. AWS SageMaker was used for model training and deployment, while AWS Glue handled ETL tasks for processing stock data. Historical data and model artifacts were stored in Amazon S3, with real-time predictions managed in AWS DynamoDB. Unstructured financial data was stored in Azure Blob Storage, and Power BI Embedded provided advanced analytics.

Python LSTM Pandas Docker AWS SageMaker AWS Glue Amazon S3 Amazon DynamoDB Azure Blob Storage Power BI Embedded

Scalable Load balancer

Built a lightweight HTTP load balancer built using Go, implementing the Weighted Least Connections (WLC) algorithm. It intelligently distributes incoming client requests across multiple backend servers based on the number of active connections and server-assigned weights. he system features real-time health checks and dynamic server pool management, ensuring optimal performance and reliability even during backend server failures or maintenance.

Go Weighted Least Connections Load Balancing Concurrency

Secure Multi-User Chat Application

Developed a highly scalable and concurrent chat server in Go, leveraging lightweight goroutines, channels, and efficient TCP socket programming to support private, public, and group messaging. Integrated and configured TLS encryption to establish secure communication channels, ensuring end-to-end data privacy and integrity between clients and the server.

Go TCP Sockets net package Goroutines Channels TLS Encryption Mutex

Multi-Producer Consumer Synchronization in Java

Collaborated in a team of three members to built a Multi-Producer-Consumer architecture using Java's concurrency primitives. The solution features a thread-safe shared buffer for coordinating multiple concurrent threads with robust synchronization, ensuring efficient thread management. Producers yield at capacity limits, while consumers wait during empty states.

Java Concurrency Multithreading Locks Semaphores

Personal Health and Fitness Application

Collaborated with a team to build an ultimate personal health and fitness tracker designed to keep users motivated, monitor their progress, and help them achieve fitness goals with ease. The platform allows users to track workouts, set personal goals, view performance analytics, and receive intelligent, interactive support through a built-in chatbot for a more engaging and personalized fitness journey.

React HTML Tailwind CSS JavaScript Node.js Express PostgreSQL MySQL RESTful API Postman Dialogflow

Voice-Activated Electronic Health Record Documentation

Developed a voice-based system to transcribe patient-physician interactions and generate structured clinical reports, optimizing physician workflow by reducing documentation time by 30%, enabling increased focus on patient care, and improving healthcare facility efficiency.

HTML CSS JavaScript SQL3 Django React Flutter Dart

Postcraft Go

Postcraft Go is a lightweight and efficient RESTful API developed in Go for seamless blog post management, supporting full CRUD operations—create, read, update, and delete. Built using the Gin web framework and GORM for ORM-based database interactions, it ensures clean and structured endpoints with JSON request/response handling and proper HTTP status codes for robust error management

Go Gin-Gonic web framework GORM ORM library for database operations Postman (for API testing) PostgreSQL REST API

Hey,I'm Aakaash M S

Software Developer • Tennis

About Me

My Projects

Think Board

Cine Verse

Subscription Tracker

Multivariate Stock Market Forecasting

Sales Forecasting and Churn Prediction

Deep Learning

Fake News Detection Using BERT

Customer Churn Analytics Dashboard

Data Driven A/B Testing

CSV-to-Data Warehouse ETL Pipeline

Hospital Data Warehouse

MultiCloud StockCast

Scalable Load balancer

Secure Multi-User Chat Application

Multi-Producer Consumer Synchronization in Java

Personal Health and Fitness Application

Voice-Activated Electronic Health Record Documentation

Postcraft Go

My Skills

Programming Languages

Frameworks and Libraries

Database

Cloud Technologies

Other Skills

My Experience

Software Developer Intern

Nutmeg Software Solutions

Contact Me

Let's work together

Email

Phone

Location