Sandeep Balabantaray

I'm an AI/ML engineer and full-stack builder. I design algorithms, ship Python backends, and turn data into products people can use — from adaptive learning and healthcare tools to legal AI. AI isn't a tagline for me; it's how I build.

Download Resume

LinkedInGitHubCodingProfile
Sandeep Balabantaray

Click to learn more

Highlights

I've built my journey around challenging myself in AI and problem-solving. I reached the Top 10 globally in Meta Hacker Cup's AI Track and was selected among the Top 15 projects in the Google × Kaggle Agents Intensive Capstone. I've also worked across four organizations, learning to adapt fast, compete hard, and deliver.

Click to flip back

Education Journey

Academic Progress & Achievements

10th Standard

Kendriya Vidyalaya No-5, Bhubaneswar, Odisha
2018 - 2019
Completed
Successfully completed secondary education with excellent performance
Percentage: 91.4%

12th Standard

Kendriya Vidyalaya No-5, Bhubaneswar, Odisha
2020 - 2021
Completed
Completed higher secondary education with outstanding results
Percentage: 94.6%

B.Tech (CSE)

VIT Bhopal University, Kothri Kalan, Madhya Pradesh
2022 - 2026
Completed
Bachelor of Technology in Computer Science Engineering — passed and awarded the provisional degree
Final CGPA: 8.74

BUILDING AI. SCALING INTELLIGENCE.

MACHINE LEARNING • GENERATIVE AI • AGENTIC SYSTEMS

TensorFlowTensorFlow
KerasKeras
Scikit-learnScikit-learn
OpenCVOpenCV
CUDACUDA
FlaskFlask
PythonPython
SQLSQL
Next.jsNext.js
TailwindTailwind
NLTKNLTK
Hugging FaceHugging Face
DjangoDjango
GeminiGemini
ReactReact
DockerDocker
AWSAWS
ChromaDBChromaDB
CloudinaryCloudinary
DeepFaceDeepFace
FastAPIFastAPI
Google ADKGoogle ADK
LangChainLangChain
LangGraphLangGraph
LangSmithLangSmith
MCPMCP
OpenRouterOpenRouter
OptunaOptuna
QdrantQdrant
SeleniumSelenium

Certifications

Professional Achievements & Recognition

Major accomplishments

Top 15Nov 2025 – Dec 2025

Google × Kaggle

Agents Intensive Capstone

Selected among the Top 15 teams. Built the face-recognition module and Gemini-powered image extraction with re-ranking for semantic retrieval at 99% accuracy.

Top 10Finalist

AI Track

Meta Hacker Cup

Reached the Top 10 globally and finished as a finalist in Meta’s official AI-focused coding competition.

Experience

PROJECTS

CodeBlaze

CodeBlaze

nextjstailwind cssdjangosqlreactjsdockergeminiAWS
LinkVideo

About

Codeblaze is a next-gen online judge platform for coders and competitive programmers. It features user profiles, problem catalogs, a Monaco-based code editor, real-time feedback via WebSockets, and a secure Docker-powered judging system.

Tech Stack
  • Frontend: React.js, Tailwind, Monaco Editor
  • Backend: Django + DRF, Django Channels (WebSockets)
  • Database: SQLite (dev), scalable to PostgreSQL/MySQL
  • Judging: Docker-sandboxed workers
  • Infra: Vercel (frontend), AWS EC2 + Nginx + SSL (backend)
CourseCraft

CourseCraft

reactjstailwind cssflaskmongodbpythonhugging face
LinkVideo

About

CourseCraft is an AI-powered learning platform that delivers personalized tutoring, adaptive learning paths, and ML-driven course recommendations. It features an interactive dashboard, real-time AI chat tutor, and secure user management with a modern responsive UI.

Tech Stack
  • Frontend: React (Vite), Tailwind, Framer Motion, Chart.js
  • Backend: Flask (Python), MongoDB Atlas, Flask-JOSE Auth
  • AI: Hugging Face (Falcon-7B, BART-MNLI) for tutoring + recommendations
  • Infra: Render (backend), Vercel (frontend), Cloud MongoDB
Walmart Bot

Walmart Bot

scikit learnreactjstailwind cssflaskpythongemini
LinkVideo

About

Walmart AI Grocery Assistant is a shopping copilot that lets users build and refine grocery lists via chat, extract items from images/PDFs with OCR, and get product recommendations through market-basket analysis. Powered by Google Gemini for natural-language understanding, it integrates Cloudinary for uploads and a Chrome extension to streamline actions on walmart.com, with sessions maintaining context and outputs returned as JSON.

Tech Stack
  • Frontend: React (Vite), Tailwind, axios, lucide-react
  • Backend: Flask (Python), Google Gemini, OCR (PyMuPDF, Tesseract, Pillow), Cloudinary
  • Recommender: Flask + mlxtend (FP-Growth/Apriori), pandas
  • Extension: Chrome MV3 (service worker, content script, popup)
  • Infra: Vercel/Netlify (frontend), Render
CourseFinder

CourseFinder

flaskpythonscikit learnNLTKtailwind cssReact
Link

About

CourseFinder helps users discover courses based on past enrollments and interests, and includes a Python-based machine learning sentiment analyzer to evaluate course reviews for better recommendations.

Tech Stack
  • Frontend: React, Tailwind
  • Backend: Flask (Python)
  • ML: Sentiment analysis with scikit-learn and NLTK
  • Infra: Render (hosted at coursecom.onrender.com)
Nexus AI Smart Gallery

Nexus AI Smart Gallery

google adkgeminifast apilangchaindeep faceqdrantchroma dbcloudinary
Link

About

Nexus AI Smart Gallery is an intelligent multi-agent gallery system that transforms your photo library into a searchable knowledge base. By combining Biometric Identity (Facenet) with Semantic Understanding (Gemini), it allows you to search for memories using complex queries like "Dad eating pizza at the beach" while automatically offloading assets to the cloud to save local storage.

Tech Stack
  • Frontend: Google ADK
  • Backend: Fast API
  • AI: Langchain, Gemini
  • Database : Qdrant, Chroma DB
  • Cloud: Cloudinary
Breast Cancer Detection IOT

Breast Cancer Detection IOT

reactjstailwind cssfast apipythonscikit learnTensorFlow.svg
Link

About

This project functions as an edge computing system where the Raspberry Pi serves as the central processing unit. Patient data is transmitted from a web interface directly to the Pi. The Pi then runs an onboard Machine Learning algorithm to analyze the received data and generate a real-time breast cancer prediction.

Tech Stack
  • Frontend: React, Tailwind
  • Backend: Fast API
  • ML: Scikit Learn, TensorFlow
Autonomous QA Agent

Autonomous QA Agent

streamlitlangchainselenium
LinkVideo

About

The Autonomous QA Agent is an intelligent testing assistant designed to streamline the software testing lifecycle. By combining a RAG pipeline with a Streamlit interface, the system builds a knowledge base from uploaded requirements (PDF/Markdown) and target HTML files. It allows users to generate comprehensive test plans and immediately convert them into runnable Python Selenium scripts, ensuring that all automation logic is strictly grounded in the provided business rules and technical specifications.

Tech Stack
  • Frontend: Streamlit
  • Backend: Fast API
  • AI: Langchain
  • Testing: Selenium
Eigen Solver Pro

Eigen Solver Pro

streamlitpythonfast api
Link

About

Developed a Python-based computational tool to calculate eigenvalues and eigenvectors, demonstrating the mathematical core of Machine Learning algorithms. The project illustrates how these components identify directions of maximum variance in high-dimensional data, simulating the mechanics of Principal Component Analysis (PCA) for dimensionality reduction and feature extraction.

Tech Stack
  • Frontend: Streamlit
  • Backend: Fast API
  • ML: Numpy