Let's Connect!
background

Hi, Shruti here.

I developIdeasIdeasPatternsPatternsModelsModelsCodeCodeIdeasIdeasPatternsPatternsModelsModelsCodeCode

into Visual Solutions

Currently based in Dehradun, Uttarakhand, India

Know me

arrow
Years of Learning
Projects Built
Internship
Research Projects

Get to know me

About me

I am passionate about building data-driven solutions at the intersection of Artificial Intelligence, Machine Learning, and real-world problem solving. My work focuses on deep learning, computer vision, and applied AI research, with a focus on sustainability and real-world applications.

Currently deepening my expertise in advanced ML techniques, data analysis, and model deployment to build scalable AI solutions.

Beyond the technical realm, I am a nemophilist who loves spending time in nature. I enjoy book reading, writing, learning about new things, capturing landscape photos, and editing them for fun.

Shruti Semwal

Technologies I've been working with

Skill Set

React.js

Next.js

SQL

Python

Tensorflow

💻 My work Overview

Experience
Research Scholar

Research Scholar

🗓️ August 2024 - March 2025

Responsibilities

  • Engineered a deep learning pipeline using Python, TensorFlow, and OpenCV to develop a U-Net-based semantic segmentation model for urban building facades, achieving 98.1% accuracy and 87% mean Intersection over Union.
  • Authored and presented a conference paper at the 4th International Conference on Innovative Sustainable Computational Technologies-2024 (CISCT) co-sponsored by IEEE UP Section.
  • Authored a thesis exploring deep learning applications, with a related journal published in Evergreen.
  • Published Patent application, demonstrating measurable air pollution levels reduction; from 500 ppm to 165 ppm, using Arduino, MQ135 sensor, and cloud-based monitoring for sustainable Urban Greening
Programmer Analyst Trainee

Programmer Analyst Trainee

🗓️ January 2022 - July 2022

Responsibilities

  • Completed an intensive ADM DotNet FSE internship, with an overall score of 87%.
  • Developed and delivered a .NET Core MVC web application in a cross-functional 5-member team (60-member cohort), meeting all sprint goals in an Agile environment.
  • Achieved a performance evaluation score of 4.2/5.
UNet with Attention-DL Model

Optimizing Green Building Implementation via Deep Learning-Powered Segmentation Techniques

Designed a U-Net model, integrating Spatial Multiplicative Cross Attention Mechanism for facade segmentation, reaching 87% mean IoU and 93% F1-Score.

SafeWalk | Pedestrian Infrastructure Safety

A computer vision pipeline built using YOLOv8 and OpenCV for pedestrian sidewalk safety analysis and obstacle detection.

Next Book Store

Create Books' themes, display them and browse other books through a clean minimalist App called Next Book Store.

Python, Pandas, Pytrends

Google Search Analysis

This project analyzes Google search trends to explore how search interest in AI-related terms has evolved over time and across regions.

Python, Statistics, Numpy

Hypothesis Testing

Applied a one-tailed t-test to determine if projected ATM transactions exceed the 4,000/month profitability threshold for a data-driven investment decision.

React, Firebase, Vercel

House Marketplace App

List, browse and filter properties for sale or for rent.

NodeJS, Express, MongoDB, Tailwind CSS, Socket.io

Online Room Decor Store

Online Room Decor Store with Real-time Order Tracking.

Have questions or ideas?

Get in Touch – Let’s Connect