Nishith Chowdary Mareddy

Build and publish applied AI, ML, GenAI, and data science projects focused on RAG applications, LLM orchestration, ML pipelines, computer vision, data workflows, and user-facing AI tools.

• Currently working on a Study Abroad AI CRM Assistant.
• Developed ClaimShieldAI, MedTranslateAI, and Jarvis-Techie AI Assistant.
• Built ML/data science projects across airfare prediction, churn analysis, classification, feature engineering, and model evaluation.
• Published source code, demos, documentation, and implementation notes on GitHub.

Supported applied data, AI, and research-oriented academic work across data preprocessing, exploratory analysis, ML experimentation, dashboarding, documentation, and technical research support.

• Built and maintained Python workflows to clean, transform, validate, and analyze structured datasets.
• Performed EDA and converted raw datasets into clear analytical summaries.
• Assisted with ML experimentation, model validation, metric comparison, and error-pattern review.
• Explored LLM, RAG, and GenAI use cases for research support and workflow automation.

Led student technical initiatives and contributed to applied research projects across IoT, machine learning, trustworthy social reviewing systems, technical seminars, and academic project showcases.

• Built an IoT-enabled smart water-level monitoring bottle using ultrasonic sensing and ESP32 wireless connectivity.
• Contributed to research on user trustworthiness assessment in social reviewing systems.
• Published research work on airfare price prediction using machine learning.
• Supported project expos, seminars, mentoring, presentation, and project coordination.

Completed a full-stack web development internship focused on responsive web applications, backend functionality, APIs, debugging, and deployment fundamentals.

Completed a front-end development internship focused on responsive web design, UI implementation, page structure, and performance-aware web development.

AI-powered insurance claims assistant using RAG, document understanding, and LLM orchestration for claims review and decision workflows.

Medical translation system for multilingual healthcare communication using LLM-powered real-time language support.

Voice-enabled agentic assistant with tool use, persistent memory, and workflow execution capabilities.

Built classification pipeline using Logistic Regression, Decision Trees & Random Forest. Applied SMOTE to fix class imbalance, boosting minority recall by 22%.

End-to-end ML pipeline across 1,814 records using XGBoost, SVM & ensemble methods. Handled 12% missing data, IQR outliers, StandardScaler normalization.

Real-time alertness system using EAR algorithm & dlib. Analyzed 5,000+ frames to calibrate thresholds. Triggers audio warning after 3-second closure.

ESP32-based IoT bottle system for real-time water level detection, wireless monitoring, and smart hydration management.

Engineered bi-directional video streaming platform. Optimized frame encoding pipeline for consistent quality across variable network conditions.

Full-stack transport management app. Normalized 8-table schema, PayPal gateway integration, role-based auth, searchable SQL interface.

Small language model-based resume optimization system focused on reducing inference cost while improving structured resume generation.