Atharv Bhosale — Data Science & Machine Learning

Project Work

Simplicity Takes, OBSESSION.

Case index01 — 07

ML Model Monitoring and Drift Detection — project visual

01 / 07 — MLOps
Model Monitoring & Drift
2026 · In progress

ML Model Monitoring & Drift Detection

An end-to-end monitoring framework guarding models in production — 25+ automated tests spanning univariate, multivariate, data-quality, data-drift and concept-drift scenarios, with eight statistical detectors from KS and Chi-Square to PSI, Wasserstein distance, Jensen–Shannon divergence, CUSUM and PCA-Mahalanobis, explained through SHAP. Served by FastAPI, surfaced in a React dashboard, shipped in Docker.

View case

04 / 07 — MLOps · Classification
Customer Churn
2026 · Shipped

IBM Customer Churn MLOps

Built an end-to-end MLOps pipeline for IBM Customer Churn Prediction using MLflow, FastAPI, Docker, GitHub Actions, and Kubernetes. Achieved 74.02% accuracy with automated testing and scalable model deployment.

View case

02 / 07 — GenAI · NLP · Vision
Emotional Intelligence
2026 · Shipped · Under review, Springer Nature DSEJ (Q1)

MindTrace AI

A real-time, multimodal emotional-intelligence system that reads text and imagery with Hugging Face Transformers and computer vision — paired with an LLM-powered chatbot for semantic analysis and deeply personalized responses. The research behind it is currently under review at Springer Nature's Data Science and Engineering Journal.

View case

03 / 07 — Applied ML
Dual-Domain Risk Intelligence
2026 · Shipped

MatRisk AI

One system, two risk languages: physics-constrained material degradation and financial risk, modelled with Scikit-Learn to R² scores of 0.43 and 0.47. Versioned data and models through DVC, validated with Pytest — a reproducible-ML backbone behind a FastAPI and Streamlit front end, charted in Plotly.

View case

06 / 07 — E-Commerce
Grocery Platform
Shipped

GroMart

A grocery store management and e-commerce platform offering an intuitive shopping experience.

View case

Experience

Data Scientist InternZettheta Algorithms Pvt. Ltd.

Developed AI and machine-learning solutions in Python — statistical analysis and data-driven modelling to extract insight from complex datasets; applied ML and statistical methods to analyse, model and evaluate risk across physics and financial domains.

Jul 2026 — Aug 2026

Research

MindTrace AIUnder review · Springer Nature DSEJ

A multimodal, real-time emotional-intelligence architecture combining NLP, transformers and computer vision with an LLM-driven semantic layer — submitted to Data Science and Engineering Journal (Impact Factor 4.2, Q1).

2026

GroMartPublished · IJCRT

Research on grocery-market intelligence and demand analytics, published in the International Journal of Creative Research Thoughts (2025–26), Impact Factor 9.17 (IJCRT metric via Google Scholar / Semantic Scholar methodology).

2025 — 26

Gallery

Milestones & Honors.

TCS World 10K finisher

Historic Run S5

Winner Spot Painting at KITCOEK

Art showcasing at KITCOEK

First Half Marathon on My 21st Birthday

Art showcasing at KITCOEK

First Research paper on IJCRT

Capabilities

ML / DL

Scikit-Learn · XGBoost
TensorFlow · PyTorch
Supervised / Unsupervised ML
Hyperparameter tuning
Feature selection & engineering
SHAP explainability
Computer vision · Transformers

AI / GenAI

LLMs · RAG pipelines
Hugging Face ecosystem
NLP & semantic analysis
Embeddings
Generative AI

MLOps & Cloud

Model monitoring & drift detection
Model evaluation & deployment
MLflow · DVC · Pytest
Docker · Git
AWS (EC2, S3) · GCP

Languages

Python
SQL
Java
C
JavaScript / TypeScript

Frameworks

FastAPI · Flask
Streamlit
React · Node.js
Plotly

Data

pandas · NumPy
EDA & data analysis
Statistical analysis & hypothesis testing
MySQL · PostgreSQL · MongoDB

About

Data Scientist by Discipline, Athlete by Grit, Artist by Eye.

Final-year Data Science student at KIT CoEK (CGPA 8.5) with real-world experience shipping ML solutions at Zettheta Algorithms. My research covers grocery demand analytics (IJCRT, IF 9.17) and deep generative systems (currently under review with Springer Nature). I care deeply about model honesty, post-deployment monitoring, and building AI that feels intuitive to humans.

The Athlete inside me brings raw endurance to long debug cycles, high-pressure execution, and a daily grind mindset that values incremental gains and continuous performance optimization.

Artist side adds visual intuition to EDA and complex data storytelling, creative problem-solving when default models fail, and human-centric empathy for the user experience.

Portrait placeholder — replace with your photograph

Contact

BEWATERMYFRIEND

atharvbb17@gmail.com GitHub ↗ LinkedIn ↗ Kaggle ↗ Download résumé ↗