AI Systems · Decision Intelligence

Nivedhaa 'Nia'
Naresh Kumar

Designing AI systems for decisions that can’t wait.

Scroll to explore
About

Undeniably
curious.

Data Science @ Boston University. Passionate about using machine learning to solve real problems where the stakes are high.

Long-term, I'm interested in applying AI in mission-critical, time-sensitive environments where rapid, reliable decision-making is essential.

Nivedhaa
Experience

Where I've worked.

TCS
Summer 2026 Incoming
AI Engineer Intern
Tata Consultancy Services · Montreal, Quebec · Hybrid
  • Incoming AI Engineer Intern focused on applied AI and machine learning solutions
Emory
May – Oct 2025
AI/ML Intern
Emory University · Atlanta, Georgia
  • Designed CNN and Transformer models in PyTorch/TensorFlow, achieving 89% accuracy on endometrial ultrasound datasets
  • Built end-to-end pipeline: preprocessing, segmentation, feature engineering, and contrastive learning
  • Cut model training time 6× (3+ hrs → 30 min) through architecture and preprocessing optimization
Georgia Tech
May – Aug 2025
Data Engineering Intern
Georgia Tech · Atlanta, Georgia
  • Built full-stack React dashboard for behavioral experiment data collection and visualization
  • Designed real-time data ingestion pipeline integrating hardware instrumentation
  • Enabled psychologists to run studies 4× faster through streamlined experiment setup
University of Rochester
Jan – Dec 2025
Research Assistant
University of Rochester · Rochester, New York
  • Queried and cleaned large-scale signal metadata using SQL for downstream statistical analysis
  • Co-designed ultrasound probe for real-time fetal brain monitoring to detect birth asphyxia
  • Generated visualizations from raw ultrasonic signals to support clinical interpretation
University of Rochester
Aug – Dec 2025
Chemistry Teaching Assistant
University of Rochester · Rochester, New York
  • Led two weekly recitation sections for 30+ students, improving exam scores by 30%
  • Graded 120+ exams per cycle with detailed performance feedback
Projects

Things I've built.

Ultrasound Image Quality Filter

Trained a modified ResNet50 classifier with AdamW + cosine LR scheduling. Applied UMAP and distance-based outlier detection to flag noisy images and visualize dataset structure.

PythonScikit-learnUMAPNumPy
View on GitHub →
Stock Financial Analysis Tool

Automated data pipeline to scrape and structure financial metrics, integrating live market data via yfinance. Rule-based valuation framework using P/E ratios, revenue growth, and earnings trends.

PythonBeautifulSoupyfinance
View on GitHub →
Cancer Detection Pipeline

End-to-end CNN/Transformer pipeline for endometrial cancer detection from ultrasound images. Achieved 89% classification accuracy using contrastive learning and custom feature engineering.

PyTorchTensorFlowCNNsTransformers
Metabolic Cost Decision-Making App

Full-stack React application built for a Georgia Tech behavioral study exploring how physical effort influences decision-making, enabling real-time experimental tasks and structured participant data collection.

ReactJavaScriptFull-Stack
View on GitHub →
Skills

What I work with.

Languages

  • Python
  • SQL
  • Rust
  • JavaScript

Machine Learning

  • PyTorch & TensorFlow
  • Scikit-learn
  • CNNs & Transformers
  • Image Segmentation
  • Deep Learning

Data & Analytics

  • Pandas & NumPy
  • Spark
  • Tableau
  • UMAP

Tools & Platforms

  • React & HTML/CSS
  • Git & Linux
Contact

Let's connect.

Open to research collaborations, internship opportunities, and interesting conversations.