Data Scientist & AI Engineer

Building
I
Solutions for Tomorrow

Transforming complex data into actionable insights and deploying state-of-the-art AI models.

View Projects Download Resume

About Me

Subhadip Hensh

Hi, I am Subhadip Hensh, an aspiring Data Scientist and AI Engineer with a strong foundation in mathematics and computer science. I am a fresher who has built skills through formal education, structured training, and extensive hands-on project work.

Earlier in my academic journey, I explored Full Stack Development but realized that application development was not aligned with my interest in analytical thinking. Mathematics had always been my strongest subject. This changed during my MSc, when I was introduced to Data Science and observed how strongly the field relies on mathematical reasoning and statistics. That realization motivated me to consciously choose Data Science and Artificial Intelligence as my career path.

I hold a BSc in Computer Science from Bankura Sammilani College (2019–2022) and an MSc in Computer and System Sciences from Visva-Bharati University (2022–2024). I have also completed advanced training in Data Science and AI with IIT Roorkee and Microsoft. Currently, I am focused on strengthening my expertise through continuous learning and hands-on experimentation, with a strong desire to grow in a professional environment.

Tech Stack

Tools and technologies I use to build intelligent systems.

PythonExpert
TensorFlowAdvanced
PyTorchAdvanced
Scikit-LearnExpert
PandasExpert
NumPyExpert
SQLIntermediate
Power BI
Power BIAdvanced
AWSIntermediate
PythonExpert
TensorFlowAdvanced
PyTorchAdvanced
Scikit-LearnExpert
PandasExpert
NumPyExpert
SQLIntermediate
Power BI
Power BIAdvanced
AWSIntermediate
PythonExpert
TensorFlowAdvanced
PyTorchAdvanced
Scikit-LearnExpert
PandasExpert
NumPyExpert
SQLIntermediate
Power BI
Power BIAdvanced
AWSIntermediate
PythonExpert
TensorFlowAdvanced
PyTorchAdvanced
Scikit-LearnExpert
PandasExpert
NumPyExpert
SQLIntermediate
Power BI
Power BIAdvanced
AWSIntermediate
Azure
AzureIntermediate
DockerIntermediate
GitAdvanced
Matplotlib
MatplotlibAdvanced
Seaborn
SeabornAdvanced
PlotlyIntermediate
SciPyAdvanced
LangChain
LangChainAdvanced
LangGraph
LangGraphIntermediate
Azure
AzureIntermediate
DockerIntermediate
GitAdvanced
Matplotlib
MatplotlibAdvanced
Seaborn
SeabornAdvanced
PlotlyIntermediate
SciPyAdvanced
LangChain
LangChainAdvanced
LangGraph
LangGraphIntermediate
Azure
AzureIntermediate
DockerIntermediate
GitAdvanced
Matplotlib
MatplotlibAdvanced
Seaborn
SeabornAdvanced
PlotlyIntermediate
SciPyAdvanced
LangChain
LangChainAdvanced
LangGraph
LangGraphIntermediate
Azure
AzureIntermediate
DockerIntermediate
GitAdvanced
Matplotlib
MatplotlibAdvanced
Seaborn
SeabornAdvanced
PlotlyIntermediate
SciPyAdvanced
LangChain
LangChainAdvanced
LangGraph
LangGraphIntermediate

Featured Projects

A selection of my work in Data Science and AI.

ResearchPilot: Production RAG System

ResearchPilot: Production RAG System

End-to-end RAG pipeline ingesting 3,500+ ArXiv papers into 300k+ semantic chunks using BGE embeddings. Features hybrid BM25/Dense retrieval with Reciprocal Rank Fusion and two-stage cross-encoder re-ranking.

PythonLangChainFastAPIQdrantHuggingFaceNext.jsDocker
GitHub Demo
FruitnetSH: Neural Plant Diagnostics

FruitnetSH: Neural Plant Diagnostics

Advanced Pomegranate Pathogen Detection utilizing neural network classification and client-side TensorFlow.js inference. Detects Alternaria, Anthracnose, Bacterial Blight, and Cercospora with sub-second latency.

PythonTensorFlowNext.jsTensorFlow.jsComputer VisionDeep Learning
GitHub Demo
Fashion MNIST Magic Blackboard

Fashion MNIST Magic Blackboard

A digital 'Magic Blackboard' for real-time fashion item classification. Built with TensorFlow.js, it uses CNN and FNN models to identify 10 types of hand-drawn apparel directly in the browser.

PythonTensorflowFNNTensorFlow.jsComputer VisionCNNDeep LearningWeb AI
GitHub Demo

Certifications & Credentials

Detailed verification of my technical expertise and professional growth.

Advanced Certification in Data Science and AI

IIT Roorkee (iHUB DivyaSampark) • 2025-26

View Credential

SQL Advanced

HackerRank • 2024

View Credential

Intermediate Machine Learning

Kaggle • 2025

View Credential

Intro to Deep Learning

Kaggle • 2026

View Credential

Problem Solving Basic

HackerRank • 2024

View Credential

Practical Experience

A showcase of hands-on expertise through end-to-end Data Science and AI projects.

Independent Project Lead

Netflix Recommendation Engine
Sep 2025 – Oct 2025
• Built a collaborative filtering system using SVD matrix factorization on the Netflix Prize Dataset (17M+ interactions). • Achieved a test RMSE of ~0.997 through rigorous quality benchmarking and 70th-percentile data filtering. • Developed an end-to-end ML pipeline from raw hierarchical data parsing to personalized ranking generation.

Data Science Capstone Developer

Walmart Sales Forecasting Project
Oct 2025 – Dec 2025
• Implemented time-series forecasting for 45 Walmart stores using SARIMA, Prophet, and stacked LSTMs with GPU acceleration. • Engineered 12-week predictive models achieving low MSE/MAE through outlier detection (IQR) and rolling statistical features. • Performed comprehensive EDA on 6,400+ records to identify high-impact sales drivers like seasonality and holiday trends.

Computer Vision Project Lead

FruitnetSH: Plant Diagnostics
Dec 2025 – Jan 2026
• Developed a neural network classification system for pomegranate pathogens (Alternaria, Anthracnose, Blight, etc.). • Integrated client-side TensorFlow.js for sub-second, real-time disease detection in the browser. • Designed and trained a custom CNN architecture to achieve high diagnostic precision across 5 distinct fruit disease classes.

AI Systems Architect

ResearchPilot (Production RAG System)
Jan 2026 – Mar 2026
• Architected a production-grade RAG pipeline ingesting 3,500+ ArXiv papers into a Qdrant vector database (300k+ chunks). • Optimized retrieval precision using Hybrid Search (BM25 + Dense) and two-stage cross-encoder re-ranking. • Deployed a full-stack solution with FastAPI backend (Docker/HuggingFace) and Next.js frontend with math rendering.

Get In Touch

Interested in collaborating or finding out more? Let's connect.

Let's Talk

I'm open to freelance opportunities and full-time roles. If you have any questions or just want to say hi, feel free to reach out!

subhadiphensh31102000@gmail.com