Portfolio / 2026

Engineering machines
that can see, measure
and make sense.

I'm Meet, a Lead Computer Vision Engineer with five years of experience moving ideas from research notebooks into reliable production pipelines.

Read the case studies →
Image registrationStereo calibrationDepth estimationModel deployment

Selected case studies

Systems with
measurable outcomes.

Four notes from production computer vision.
01

Perception pipeline

Sharper scans, faster mosaics

Optimized shift detection, image registration and mosaic generation for DepScan, improving visual sharpness while cutting generation latency by 30%.

Image registrationMosaicingC++
30%latency reduction
02

Few-shot recognition

Learning a new vehicle from ten examples

Combined deep feature detection, matching and an incremental few-shot classifier to identify new vehicle categories with limited training data.

PyTorchFeature matchingFew-shot
90%detection / 10 samples
03

Robust anomaly detection

Fewer false alarms in changing conditions

Used one-class learning, SAM-based background removal and domain adaptation to make change detection more dependable outside controlled datasets.

Anomaly detectionSAMDomain adaptation
−25%false positives
04

Deployment systems

From model file to production service

Built an automated calibration utility and a framework-agnostic Python inference server to make multi-model delivery faster and repeatable.

PythonCUDAModel serving
5×faster deployment

Profile

A research instinct,
with a product bias.

For five years, I've built computer vision systems for under-vehicle scanning: registration, stereo geometry, depth estimation, anomaly detection and the software that carries those models into production.

I care about the complete path from an experiment to a maintainable system, including data, performance, deployment and the people who operate it.

Lead Computer Vision Engineer

Vehant Technologies / DepScan

Teaching Assistant

DA-IICT / Computer Architecture & Linux

M.Tech., Software Systems

DA-IICT

B.Tech., Computer Engineering

Ganpat University

Publications

2023

DepScan: 3D Under Vehicle Scanning System

AIML Systems

↗2020

Segregating and Recognizing Human Actions from Video Footages Using LRCN Technique

AMLTA

↗