Yugandhar Patil

Adaptive RAG Portfolio Chatbot

An agentic RAG chatbot that intelligently routes every query to the right pipeline — documents, web search, or general LLM — powered entirely by a local model. No OpenAI. No API costs. 100% free.

Multi-Agent Code Review System

4 specialized AI agents (Security, Performance, Code Quality, Test Coverage) analyze GitHub PRs in parallel, then a meta-agent synthesizes findings into a single prioritized review. Built with Google Gemini structured JSON output and asyncio for true parallel execution.

Student Performance Prediction

End-to-end ML pipeline predicting student pass/fail and final grade (G3) using demographic, academic, and behavioral data. Implemented Logistic Regression, Decision Trees, Random Forest, and SVM with full preprocessing and cross-validation.

Graduate Admission Predictor

Regression model predicting university admission chances based on GRE, TOEFL, CGPA, and other profile features. Used GridSearchCV for model selection across multiple algorithms.

Spam Email Detector

NLP model classifying spam emails using TF-IDF vectorization and logistic regression. Applied full text preprocessing pipeline including tokenization, stopword removal, and feature extraction.

Uber Ride Cancellation Analysis

EDA on 150,000 Uber ride bookings to uncover cancellation patterns across drivers, customers, vehicle types, pickup/drop locations, and time-of-day. Surfaced actionable insights to improve ride fulfillment rates.

Frequent Pattern Mining Engine

Implements and compares three frequent itemset mining approaches — Brute Force (from scratch), Apriori, and FP-Growth — on five retail datasets. Configurable min-support and min-confidence via CLI.

SecureCalc: Full-Stack FastAPI App

Full-stack web app with BREAD calculator operations, JWT authentication, and a user profile/password management system. Backed by PostgreSQL, tested with Playwright E2E, and shipped via GitHub Actions CI/CD to Docker Hub.

Binary Classification: Heart Failure

Trains and compares Random Forest, SVM (RBF), and GRU on the UCI Heart Failure dataset using 10-fold stratified cross-validation. Evaluates across 9 metrics including ROC/AUC, F1, TSS, HSS, and Brier Score.

Enhanced CLI Calculator

Command-line calculator using Factory, Memento (undo/redo), and Observer design patterns. Auto-saves history to CSV via pandas, .env-driven config, color-coded output, and 90%+ test coverage with GitHub Actions CI.

Sentiment Analysis with NLTK

Naive Bayes sentiment classifier trained on NLTK's movie_reviews dataset. Handles text cleaning, tokenization, and stopword removal to predict positive/negative sentiment on new inputs.

Image Classifier: CNN on MNIST

Convolutional Neural Network built with TensorFlow/Keras to classify handwritten digits (0–9) from the MNIST dataset. Three conv layers with ReLU activations, Adam optimizer, and one-hot encoded labels.

Google Data Analytics Professional Certificate

Google Career Certificates

8-course program covering data cleaning, analysis, and visualization using SQL, R, Tableau, and Spreadsheets. Includes a capstone project analyzing real-world datasets.

Foundations of Data Science

Google Career Certificates

Foundational course covering the data science workflow, Python basics, exploratory data analysis, and communicating insights to stakeholders.

Crash Course on Python

Google Career Certificates

Hands-on Python course covering variables, data structures, functions, OOP, and automation scripting. Designed for learners with no prior programming experience.

Claude 101

Anthropic

Introduction to Claude's capabilities, safety principles, and responsible AI usage. Covers prompt design, model behavior, and best practices for interacting with large language models.

Claude Code 101

Anthropic

Fundamentals of using Claude Code as an AI-powered coding assistant. Covers agentic workflows, code generation, debugging with Claude, and integrating Claude Code into development pipelines.

Claude Code in Action

Anthropic

Applied course building real projects with Claude Code. Covers multi-file editing, test generation, refactoring, and using Claude as an autonomous coding agent on complex tasks.

AI Fluency: Framework & Foundations

Anthropic

Framework for understanding generative AI systems, their capabilities, limitations, and societal impact. Covers responsible deployment, AI safety concepts, and evaluating model outputs.

AI Fluency for Students

Anthropic

Student-focused curriculum on leveraging AI tools effectively in academic and professional settings. Covers prompt engineering, critical evaluation of AI outputs, and ethical AI usage.

Master of Science in Computer Science

New Jersey Institute of Technology

Newark, NJ

Bachelor of Engineering in Electronics & Telecommunication

Army Institute of Technology

Pune, India