Building AI through applied projects.
A portfolio documenting my progression from data analysis and classical machine learning to neural networks, computer vision, generative AI and production-oriented deployment.
Applied AI & Data Science projects
Data → ML → AI → Deployment
PORTFOLIO
Applied AI projects.
The portfolio demonstrates a progressive learning journey across data analytics, machine learning, deep learning, computer vision, generative AI, API development and production deployment.
DEPLOYED ML
SuperKart
SuperKart is a retail forecasting project focused on predicting quarterly sales revenue across supermarkets and food marts located in different city tiers.
The predictive solution supports inventory optimization and regional sales strategy while demonstrating how a machine-learning model can be operationalized and deployed at scale.
Moving from exploratory analysis and model optimization to API development, containerization, application development and cloud-oriented deployment.
COMPUTER VISION
HelmNet
A computer-vision project designed to automatically determine whether workers are wearing safety helmets.
The objective is to support workplace safety, increase monitoring efficiency and reduce the human error associated with manual safety compliance inspections.
How augmentation, transfer learning and fine-tuning can improve generalization when working with limited image datasets.
RAG · LLM
Medical Assistant
A Retrieval-Augmented Generation architecture designed to retrieve relevant information from established medical manuals and provide context-aware responses.
The project explores generative AI for information retrieval, structured decision support and consistent access to medical knowledge.
Retrieval pipelines, grounding model responses in source material, prompt design, reliability and traceability in healthcare-related AI systems.
DEEP LEARNING
ReneWind
Predicting wind-turbine generator failures using sensor data.
The objective is to support proactive maintenance, reduce repair costs and minimize unplanned downtime by identifying potential equipment failures before breakdowns occur.
Aligning model evaluation with business consequences, especially when false negatives can result in costly equipment failures.
ENSEMBLE LEARNING
EasyVisa
Advanced machine-learning techniques applied to historical visa application data to identify factors associated with certification outcomes.
Comparing ensemble approaches, tuning model parameters, interpreting feature importance and translating model outputs into useful analytical insights.
CLASSIFICATION
Personal Loan Campaign
Customer analytics and predictive modeling for identifying individuals with a high likelihood of accepting a personal-loan offer.
Evaluating classification performance, identifying important customer characteristics and connecting analytical findings with marketing recommendations.
DATA ANALYSIS
FoodHub
Exploratory data analysis examining customer orders, restaurant demand, cuisine preferences, ratings and delivery patterns.
The objective is to transform transactional data into actionable insights that can improve customer experience and business decisions.
Data manipulation, exploratory analysis, statistical examination and communicating analytical findings through visualizations.
LEARNING JOURNEY
From data analysis to deployed AI.
CAPABILITIES
Capabilities demonstrated.
Data Science
Python
Pandas & NumPy
Exploratory Data Analysis
Data Visualization
Feature Engineering
Machine Learning
Classification
Decision Trees
Ensemble Models
Hyperparameter Tuning
Model Evaluation
Deep Learning & AI
Neural Networks
Computer Vision
CNNs
Transfer Learning
RAG & LLMs
Deployment
Flask
REST APIs
Docker
Streamlit
Hugging Face
RESEARCH · AI · UX
Building intelligent systems around real human needs.
My broader interest lies at the intersection of Artificial Intelligence, User Experience, Usability and applied software development.
Explore my research →