Sachin Maurya - Portfolio

About Me

I'm a B.Tech. (Hons) CSE student specializing in Artificial Intelligence at Chhattisgarh Swami Vivekanand Technical University (CGPA: 7.5). With hands-on experience in data science and machine learning, I focus on predictive maintenance, generative modeling, and medical image analysis. My work includes developing a digital twin for centrifugal pumps and a CNN-based MRI classifier for brain tumor detection. Proficient in Python, TensorFlow, and end-to-end ML pipelines, I aim to create impactful AI solutions.

Education

B.Tech. CSE (AI), CSVTU, Bhilai

2022 - Present | CGPA: 7.5
12th, Holy Angels Public School (CBSE)

2021 | 75%
10th, S.B.T Public School (CBSE)

2019 | 80%

Skills

Programming & Libraries

Python (NumPy, Pandas, Matplotlib, Seaborn, Scikit-learn, TensorFlow, PyTorch, OpenCV), SQL

AI/ML Core Concepts

Statistics, Data Structures, ML (SVM, RF, XGBoost), Deep Learning (CNN, LSTM, GANs), NLP

Advanced AI Technologies

LLMs, RAG, Prompt Engineering, LangChain Agents, FAISS, ChromaDB, Agentic AI

Data Handling & Processing

Data Cleaning, EDA, Feature Engineering, Web Scraping, Visualization

Tools & Platforms

Google Colab, Git, MATLAB, Tableau, VS Code, Jupyter

Cloud & Deployment

Google Cloud (Firebase), AWS (S3, EC2), Docker (basic)

Experience

Digital Twin of Centrifugal Pump (NSTL-DRDO) | IIT Roorkee

2025

Collaborated on predictive maintenance using AI/ML techniques.
Trained models with Random Forest, XGBoost, and LSTM for time-series forecasting.
Developed a live simulation dashboard for real-time pump performance and failure prediction.
Reduced unplanned downtime in defense applications.

Data Scientist | Quick CURD

2025

Framed real-world ML problems and mastered the end-to-end ML workflow.
Focused on feature clarity, data consistency, and domain alignment.
Developed scalable AI solutions for dynamic use cases.

Projects

Vibration Data Analysis of a Centrifugal Pump

Processed real-time vibration signals, applied FFT and ML models (RF, XGBoost, LSTM) for predictive maintenance and anomaly detection. Built dashboards for actionable insights.

View on GitHub

Human Activity Recognition (HAR)

Implemented ML (Logistic Regression, SVM, RF) and DL models on UCI HAR dataset. Achieved 96.7% accuracy with Linear SVC, outperforming baseline methods.

View on GitHub

Tooth Abnormality Detection

Developed YOLOv8-based detection for dental anomalies (decay, black stains). Trained on 244 images, achieving mAP50 of 82.2% (decay) and 57.8% (black stains).

View on GitHub

Breast Cancer Detection

Built a classification model using ML/DL techniques to detect breast cancer from medical datasets. Achieved high accuracy and provided explainable insights for clinical use.

View on GitHub

Remaining Useful Life (RUL) of Pump Components

Developed ML algorithms to estimate RUL of centrifugal pump components with real-time health monitoring.

View on GitHub

Human Induced Load (GAN)

Designed a GAN-based model to calculate human-induced load using W-GAN and W-GAN-GP for stability.

View on GitHub

Brain Cancer Detection

Developed a CNN-based model for brain cancer detection using MRI images, achieving 99.5% accuracy.

View on GitHub

Multimodal AI Diagnostic Assistant

Integrated CT and X-ray image models for detecting chest and abdominal conditions using CNNs. Enabled real-time diagnostic dashboard with explainable AI.

View on GitHub

LLM-Powered Medical Report Generator

Built an agentic AI system using RAG pipelines and LLMs to generate clinical reports based on diagnostic outputs with verified medical literature context.

View on GitHub

AI-Powered Stock Trend Prediction

Developed an LSTM-based model for multi-stock forecasting. Integrated LLMs for news sentiment analysis using a RAG pipeline for better market insights.

View on GitHub