Harish Balaji

About Me

A New York University master's student working with the VIDA Lab on multi-sensor fusion and computer vision applications for traffic safety and urban mobility. In summer 2025, I contributed to the University of California, OSPO as part of Google Summer of Code, developing privacy metrics and evaluation tools for machine learning systems.

The question I keep coming back to: "Why build software if it doesn't truly matter to anyone?"

I care about building tools that people can genuinely trust and rely on. Whether it's designing privacy-first systems or developing software for sensors that help prevent accidents before they happen, I'm driven by problems that create real-world impact.

Contact Details

Email:
harish.balaji.b@nyu.edu
Location:
New York, NY

Resume

Education Details

  New York University
  GPA: 4.0/4.0
  Master's in Computer Engineering
  SASTRA University
  Bachelor's in Computer Science and Engineering
  Specialization in Cyber Security

Professional Experience

Google Summer of Code – Open Source Contributor

University of California, OSPO • Jun 2025 – Sep 2025

•Worked with a professor and researcher from Lawrence Berkeley National Laboratory (LBNL) to design privacy metrics quantifying re-identification risk across structured datasets, including a normalized risk scoring system.
•Integrated image dataset support with pre-trained vision models (CLIP, ViT, DINOv2).
•Enhanced evaluation pipeline for multi-modality and visual dashboards of quality, fairness, and privacy metrics.

Graduate Research Assistant – VIDA Lab

New York University • Jan 2025 – Present

•This work focuses on pose detection and synchronizing multimodal data (audio and video), ensuring 4K videos are synchronized at 30FPS with frame dropping detection to provide comprehensive insights into traffic safety scenarios.
•Collected data in multiple intersections of Brooklyn around 100s of GB and analyzed pedestrians of various diversity.
•Deployed HRNet-based pose estimation on traffic video feeds (1,000+ pedestrians, safety risk analysis).
•Built ML pipelines to classify 5,000+ objects and compute pedestrian speed with 95%+ accuracy.

PROJECTS

El Silencio Acoustic Explorer

Edge AI system for biodiversity monitoring using multi-branch ML pipeline with real-time audio processing. The system processes acoustic data from field sensors to identify species and monitor ecosystem health, demonstrating the potential of edge computing for environmental conservation.

RouteWise

Intelligent route optimization engine with five algorithmic solvers and smart preference scoring. The system combines multiple optimization algorithms to provide personalized routing solutions, considering factors like traffic, weather, and user preferences for optimal navigation.

DailyPod

Automated news-to-audio intelligence platform with multilingual aggregation and WhatsApp delivery. The system processes news articles from multiple sources, converts them to audio using text-to-speech, and delivers personalized news summaries via WhatsApp, making information accessible to diverse audiences.

XGChurn

Predictive analytics platform for customer churn with XGBoost classifier and SHAP interpretability. The system analyzes customer behavior patterns to predict likelihood of churn, providing actionable insights for customer retention strategies with explainable AI techniques.

BLOGS

AIDRIN Privacy-Centric Enhancements: Backend & UX Upgrades

Building privacy-first AI data pipelines with enhanced backend infrastructure, privacy metrics integration, and improved UX for AIDRIN • July 25, 2025

Improving AI Data Pipelines in AIDRIN: A Privacy-Centric and Multimodal Expansion

Documenting my GSoC 2025 journey with AIDRIN, focusing on privacy metrics and multimodal dataset evaluation for responsible AI • June 25, 2025

Hi!

I'm Harish Balaji

A Computer Engineering master's student at New York University, building scalable software and privacy-first AI solutions that transform complex data and research into impactful, real-world innovations.