Yash Shah
BengaluruIndia

Yash Shah

SDE 2 at NetApp.

I build computer-vision and machine-learning systems — from published research to the factory floor.

Selected work

DocAuth2023

Detects and localises document forgeries at 97% accuracy.

  • metadata analysis
  • Fourier analysis
  • Error Level Analysis
  • template matching
  • vector database
Effimix2022

A feature-fusion CNN for gastrointestinal disease classification, published in Diagnostics.

  • EfficientNet-B0
  • squeeze-and-excitation
  • feature fusion
  • HyperKvasir dataset
Urban-Sat2023

Segments roads and buildings from satellite imagery for urban planning.

  • U-Net
  • semantic segmentation
  • satellite imagery

All projects

Experience

  1. NetAppSDE 2

    current · Bengaluru, India

  2. Defect ScannerMachine Learning Intern

    May 2023 · Seoul, South Korea

    • Stood up a live video-streaming server on Microsoft Azure for global access, cutting streaming latency to under one second.
    • Flagged product defects from a live conveyor-belt feed on the factory floor, replacing roughly 12 hours of manual monitoring a day.
    • Led an incremental-learning model that detects and registers object classes new to the system on production lines.
    • Built an action-recognition model to reduce assembly-line defects.
  3. Bhabha Atomic Research CentreProject Intern

    May 2023 – July 2023 · Mumbai, India

    • Led an automated vehicle detection, tracking and logging system.
    • Applied ML-Ops to move level-0 systems to level-1: automated testing and monitoring plus a CI runner on a development machine, saving two hours of manual testing per run.
    • Introduced an SSIM-based performance monitor to catch failure cases of computer-vision models on video use-cases.
    • Contributed to two research articles documenting the algorithms from this work.
  4. Naldeo GroupSolar Project Consultant

    October 2022 – February 2023 · Lyon, France

    • Led the team forecasting congestion in solar grids.
    • Converted daily readings into hourly distributions, then predicted real-time grid congestion with predictive modelling to enable efficient load setup.

Publications & patents

  • A Novel Multi-Feature Fusion Method for Classification of Gastrointestinal Diseases Using Endoscopy Images

    Ramamurthy K, George T T, Shah Y, Sasidhar P

    Diagnostics 12(10), 2316, 2022 · doi

  • Method and System for Workflow Analysis

    US patent application, 2025

    App. 19/280,291 · published July 2025

Achievements

  • 2026Winner, NetApp Intern ShowcaseEngineering track
  • 2025Silver MedalistVellore Institute of Technology
  • 2025Best Outgoing StudentVellore Institute of Technology
  • 2023First place, Smart India Hackathon
  • 2023Outstanding Performance certificateBhabha Atomic Research Centre
  • Top 100 of 50,000+ teams, Amazon ML Challenge
  • 2020–23Rank 1, four consecutive academic yearsVellore Institute of Technology

Education

Vellore Institute of Technology

Integrated M.Tech, Computer Science and Engineering with Business Analytics

2020–2025 · Chennai, India · CGPA 9.52 / 10

Toolkit

Languages
Python · C++ · Java · C · R · JavaScript
ML & CV
Deep learning · Transfer learning · Computer vision · NLP & text processing · AutoML · Predictive modelling
Data
MySQL · MongoDB · Cassandra · Hadoop · Spark · Kafka · Flink
Platforms
Microsoft Azure · ML-Ops / CI · Big data · Blockchain