Selected Work · AI Portfolio · Research and Prototypes

Applied machine learning projects with a clear bias toward practical outcomes.

This collection highlights the academic and experimental side of Gaurav’s profile: deep learning, NLP, classical machine learning, data pipelines, algorithms, and research-led problem solving.

While the homepage focuses on production engineering impact, this page shows the analytical range behind it, including the IEEE-recognized food image classification research.

The standout project in the portfolio

IEEE 2023

CNN Based Study of Improvised Food Image Classification

Capstone research project exploring food image classification through multiple deep learning architectures, strong augmentation strategy, and preprocessing improvements such as background removal with U2Net.

InceptionResNetV2 ResNet50 DenseNet169 WiSER
Recognized as Best Research Presentation at the 2023 IEEE Annual Computing and Communication Workshop.

It is a strong proof point that the profile is not only implementation-heavy but also research-capable and presentation-strong.

Representative work across learning tracks

Grouped by the type of problem each project tackled, making the portfolio easier to scan than a long chronological list.

Deep Learning

Flower Species Classification and Sarcasm Detection

Combined transfer-learning based image classification with NLP experimentation for sarcasm detection, showing range across both vision and text tasks.

CNN Transfer Learning NLP
Deep Learning

Signal Quality Prediction for Communication Equipment

Modeled signal quality with neural networks and a full workflow covering exploratory analysis, preprocessing, training, and evaluation.

Neural Networks EDA Optimization
Machine Learning

Bank Loan Defaulter Prediction

Built a business-focused ML pipeline with feature selection, imbalance handling, and tuning workflows aimed at improving risk prediction quality.

Classification Feature Engineering Model Tuning
Machine Learning

Google Store App Rating Prediction

Explored supervised models and interpretable insights that could help app teams understand what drives stronger user ratings.

Decision Trees Ensembles Product Insight
Machine Learning

E-commerce Customer Segmentation

Used unsupervised techniques and PCA to group customers into actionable segments for marketing and retention analysis.

Clustering PCA Segmentation
Machine Learning

Telecom Customer Churn Prediction

Compared classical classification approaches for churn detection and translated model performance into customer retention strategy ideas.

Logistic Regression KNN Naive Bayes
Data Systems

E-commerce Reports with CQL

Designed Cassandra-oriented schemas and reporting queries for analytical exploration in a NoSQL context.

Cassandra CQL Reporting
Databases

Travego Travelers SQL Project

Structured relational data and query logic around a travel use case, demonstrating practical SQL modeling and reporting skills.

MySQL Schema Design Queries
Algorithms

City Routes, Power Grids, and Data Structures Exercises

Worked through graph traversal and linked-list based system modeling in Python, grounding the portfolio in core CS fundamentals as well as ML topics.

BFS DFS OOP

Production engineering impact lives on the main profile page.

If you want the professional story, platform work, leadership highlights, and current role at Oracle, the homepage ties those pieces together.