Computer Vision Researcher · University of Strathclyde

Hi, I'm Devjyoti.

Making sense of the world with code and curiosity! Think of my name as a blend of 'dev' for developer and 'jyoti', a spark of innovation in Sanskrit. I research, innovate and explore deep learning for agriculture, satellites and signals at the CIDCOM and CeSIP Labs.

Email me Google Scholar

CIDCOM team at the 2025 R&KE Impact Awards

Award · 2025

Strategic Themes Research & Knowledge Exchange (R&KE) Impact Award Winners

Centre for Intelligent Dynamic Communications (CIDCOM)

Blog

3D-Genesis

Reconstructing real-world objects from parametric primitives

Satellite vegetation map of agricultural fields

Blog

Satellite Vision

Leveraging AI and satellite data

News

Jul 2025Published the Alan Turing Institute Data Study Group final report for Transport for London.
2025CIDCOM won a 2025 Strategic Themes R&KE Impact Award.
2025Started work on FLORA-SAGE, federated learning for grassland biodiversity from Sentinel satellite data.
Mar 2025Paper on object detection for RF signal classification at IEEE NEWCAS.

Projects

Satellite image of agricultural field parcels

2025 – 2027Computer VisionRemote Sensing

FLORA-SAGE

Federated Learning Optimised for Remote Assessment of Species in Agricultural Grassland Ecosystems

Leverages Sentinel satellite data and diffusion models to create high spatio-temporal datasets, enabling deep learning models to identify plant and soil biodiversity in agricultural grasslands.

Depth camera view of a cow being tracked

2024 – 2027Computer Vision

Digital Dairy Chain (DDC) · UKRI

Deep learning and computer vision architectures for cattle recognition, 3D object tracking and multi-camera handoff, enhancing automated precision livestock management.

2024Computer Vision

Transport for London (TfL)

Point cloud data from Amberg trolley scans, converted into depth maps, to automatically detect and classify track features for efficient maintenance and renewal planning.

2023RF & Edge AI

Machine Learning for Spectral Analysis on SoC

A CNN-based model for real-time RF signal detection on the AMD-Xilinx RFSoC platform, optimised through quantization analysis with Vitis AI.

2020 – 2022Software Engineering

Management of Google Cloud Platform

Technical Executive

Optimised GCP performance across 300+ customer accounts monthly, improving database recovery speed and strengthening cloud security against rogue admin risks.

Publications

Data Study Group Final Report: Transport for London

The Alan Turing Institute · Data Study Group · Jul 2025

Report
Object Detection for RF Signal Classification: A Comparative Study for Spectrogram-Based Detection in RFML Applications

IEEE NEWCAS · MSc Thesis · Mar 2025

Conference
Converging Ethics and Sustainability in Engineering Through Collective Intelligence

EPC Poster · IET · Jun 2023

Poster
Multi-band multi-location RF power survey and ambient RF energy harvesting circuit

IEEE ANTS · BEng Thesis · Dec 2020

Paper
Online Personal Health Monitoring System

Global Wellness Summit · BEng Thesis · May 2020

Summit