Sandeep Balleda

Hyderabad,India

sandeep.balleda007@gmail.com

About

I am a Computer Science student and aspiring Software Developer with a strong interest in Frontend Development, Artificial Intelligence, and Full-Stack Development. I have hands-on experience working with technologies such as JavaScript, React, HTML, CSS, Python, Flask, SQL, and AI/ML tools. I enjoy building practical, user-friendly applications and exploring technologies such as RAG, AI chatbots, and real-time computer vision. I am a quick learner, problem solver, and enthusiastic team player who is continuously improving my technical and development skills. My goal is to contribute to innovative projects while growing as a professional software developer.

Skills

Python

C++

Java

JavaScript

HTML5

CSS3

SQL

React.js

Flask

Tailwind CSS

MySQL

SQLite

Machine Learning

Computer Vision

OpenCV

MediaPipe

NumPy

Pandas

RAG

AI Chatbot Development

LangChain

REST APIs

Git

GitHub

VS Code

Streamlit

huggingface

n8n

Experience

INTERN

Viswam.ai

May 2026 to Jul 2026

- Developed an AI-powered chatbot using RAG, LangChain, FAISS, and OpenAI APIs to generate context-aware responses.
- Implemented document embedding and vector-based similarity search to retrieve the top 3 relevant text chunks.
- Integrated an LLM with retrieved context to improve the relevance and accuracy of chatbot responses.
- Troubleshot Python dependencies, vector-store configuration, and API integration issues to improve pipeline reliability.

Projects

AirGap Parse

AirGap Parse is an offline-first, CPU-first document parsing application for converting unstructured invoices, receipts, and similar business documents into structured JSON and local SQLite records. The application is designed for air-gapped or privacy-sensitive environments. Document processing runs locally using Tesseract OCR, a local parsing pipeline, and SQLite storage with full-text search support. No internet connection is required to upload, process, review, or search documents once the local environment is set up.

Python 3.10+

FastAPI

Uvicorn

Pydantic

pytesseract

SQLite3

python-multipart

Tesseract OCR

React

DeepFake Shield: Real-Time Audio/Video Authenticator

DeepFake Shield is a sophisticated browser extension and AI-powered backend designed to combat the rise of hyper-realistic AI-generated media. It provides users with a real-time "Trust Score" to help distinguish between authentic human content and AI-generated deepfakes.

Frontend: Chrome Extension (Manifest V3

Vanilla JS

CSS3) Backend: Python (FastAPI

Uvicorn) AI/ML Layer: Audio: Hugging Face (Wav2Vec2 / Specialized Deepfake Detectors) Video: OpenCV

Dlib (Facial Landmark Consistency) Processing: Librosa (MFCC extraction)

NumPy

PyTorch

RAG Chatbot using LangChain + FAISS

The chatbot retrieves relevant document chunks from uploaded PDFs and generates accurate answers using an LLM.

Python

LangChain

FAISS

Vector Database

HuggingFace

Embeddings

OpenAI GPT Models

Streamlit UI

Medicare AI — Intelligent Medication Reminder Dashboard

Medicare AI is an intelligent medication management platform designed to help people organize their medicine schedules, receive timely reminders, track missed doses, and monitor medication adherence. By combining AI agents, voice interaction, persistent memory, and automated notifications, the platform aims to make medication management simpler, safer, and more accessible.

React.js

Tailwind CSS

Python FastAPI

LangGraph

OpenAI API

SQLite/PostgreSQL

APScheduler

and OCR for prescription extraction.

Real-Time Hand Gesture Recognition System

As modern computer interfaces are evolving into touch and voice commands, a real-time hand gesture recognition offers a natural way to control computers without physical contact. This project uses 21 hand landmarks detected by webcam through MediaPipe Hands to perform gesture classification with a machine learning model and convert them into keyboard and scroll commands on your computer.

Python

OpenCV

MediaPipe

NumPy

Machine Learning

CSV.

Janmabhoomi Civic Desk

Janmabhoomi Civic Desk is a civic-services prototype for searching village land records, submitting public-service requests, and helping administrators manage local records with a lightweight JSON data store.

Python 3.11+

Streamlit

Node.js 20+

HTML

CSS

and JavaScript JSON file storage

Education

B.tech

ICFAI University

Aug 2024 to May 2028

Certifications

Data Processing and Visualisation