Sibashis Patnaik — Data Scientist

Turning real-world datasets into production-ready AI solutions. Hands-on experience in ML, deep learning, and full-stack deployment.

View GitHub

Meet Sibashis Patnaik

Professional Summary

Data Scientist with AI/ML internship experience and full-stack web development background. Proficient in Python, SQL, and scikit-learn. Skilled at deploying predictive apps using Flask and Streamlit.

Passionate about Generative AI, RAG, and Large Language Models.

Get In Touch

📍 Berhampur, Odisha, India

📧 [email protected]

📞 +91 7978467423

About & Skills

My software engineering background gives me a distinct advantage. I don't just build ML models — I deploy and integrate them into production-ready applications.

🐍 Programming

Python, SQL

🤖 ML & Deep Learning

Scikit-learn, XGBoost, TensorFlow, Keras, PyTorch, CNN, LSTM, Transformers

📊 Data & Viz

Pandas, NumPy, Matplotlib, Seaborn

🧠 Gen AI & LLMs

LangChain, LlamaIndex, RAG, Hugging Face, OpenAI API, Vector Embeddings

🌐 Backend & DB

Flask, Streamlit, Node.js, REST APIs, MySQL, PostgreSQL, MongoDB

🚀 Deploy & Tools

Docker, Vercel, Render, Git, GitHub

Certifications

  • Professional Certificate Program in Generative AI and Machine Learning - Simplilearn (July 2026)

Academic Background

  • B.Tech — Instrumentation & Electronics, CET Bhubaneswar (2021) — CGPA 8.2
  • Diploma — Mechatronics, CTTC Bhubaneswar (2018) — 65%
Featured Project 01

SONAR-60 — Rock vs. Mine Classifier

Classifies sonar signals as rock or mine using the UCI Connectionist Bench dataset. A Logistic Regression model processes 60 frequency-band features. A Flask REST API serves predictions to a custom instrument-panel frontend. A live Canvas strip-chart renders the 60-band waveform in real time.

Stack

  • Python, scikit-learn (Logistic Regression), joblib
  • Flask REST API, Flask-CORS
  • Vanilla HTML/CSS/JS, Canvas API
  • Vercel (serverless Python + static hosting)
Featured Project 02

Rainfall Prediction Station

Estimates rainfall (mm) from live atmospheric readings. Users adjust six meteorological inputs — temperature, dew point, humidity, sea-level pressure, visibility, and wind. The app returns a real-time prediction with an uncertainty range and severity category. No more static outputs buried in a Jupyter cell.

Stack

  • Python, scikit-learn (Random Forest, 500 estimators)
  • Flask REST API, joblib
  • HTML/CSS/JavaScript (instrument-panel UI)
  • Vercel serverless deployment
Featured Project 03

SIGNAL — Telecom Customer Churn Diagnostic Console

Predicts telecom subscriber churn using the IBM Telco dataset (7,043 customers, 19 features). Three classifiers were compared via 5-fold cross-validation on SMOTE-balanced data. Random Forest achieved ~78% test accuracy. A custom "diagnostic console" frontend features an animated churn-risk gauge and feature-importance breakdown. Non-technical retention teams can score subscribers and understand why the model flagged them.

ML & Data

Python, pandas, scikit-learn (Random Forest, Decision Tree), XGBoost, imbalanced-learn (SMOTE)

Backend

Flask, Flask-CORS, REST API (/api/predict, /api/schema, /api/health)

Frontend

HTML5, CSS3, vanilla JavaScript, SVG-based animated gauge

Deploy & Tooling

Vercel (serverless), pickle serialization, pinned dependencies

Additional Projects

Diabetes Pulse App

SVM classifier on the PIMA Indians Diabetes dataset (768 records, 8 features). Achieved 76.6% test accuracy and ROC-AUC = 0.82. Flask REST API with custom real-time visualization frontend. Deployed on Vercel.

Tools: Python, scikit-learn, pandas, NumPy, Flask, Flask-CORS, HTML/CSS/JS, Vercel

BengaluruEstimate — Real Estate Price Predictor

Gradient Boosting on 8,000 listings across 30 locations. R² = 0.97, MAE = 5.69 Lakhs. Outlier rules improved R² from 0.81 → 0.97. Benchmarked 5 models via 5-fold CV. Dockerized Flask API on Render.

Tools: Python, Pandas, NumPy, Scikit-Learn, Flask, Gunicorn, Docker, Render

Experience & Education

1

AI/ML Intern — Quadsync Tech Solutions

July 2026 – Current

2

Associate Software Engineer — Fission Labs

May 2022 – Jan 2024

3

B.Tech — Instrumentation & Electronics Engineering

College of Engineering & Technology, Bhubaneswar | 2018–2021 | CGPA: 8.2

4

Diploma — Mechatronics

Central Tool Room & Training Center, Bhubaneswar | 2015–2018 | 65%

0.973

R² Score

Real Estate Price Prediction

76.6%

Test Accuracy

Diabetes Prediction (SVM)

0.82

ROC-AUC

Diabetes Prediction Model

78%

Test Accuracy

Customer Churn (Random Forest)

Let's Connect

Open to data science, ML engineering, and AI roles. Available for remote and on-site opportunities.

Phone

+91 7978467423

LinkedIn

linkedin.com/in/sibashispatnaik2000

GitHub

github.com/Sibashis216

Ready to Turn Data Into Intelligence

"I don't just build models — I build solutions that deliver measurable value."

Explore my projects, review my code, and let's discuss how I can contribute to your data team.

© 2025 Sibashis Patnaik · Data Scientist · Berhampur, Odisha, India

GitHub · LinkedIn · Email

Made with