ARCHITECTURE DESIGNER — WRITER — RESEARCHER — ENTERPRENEUR

Umar
Farooq

Computer Vision Researcher

I move ideas from peer-reviewed research into systems that run in production. Six years split evenly between two questions: how do you make a model provably better, and how do you make it survive contact with the real world.

6+ YRS EXPERIENCE
Umar Farooq
About

Two tracks, one engineer.

I'm a Computer Vision Researcher and AI Engineer working at the Hong Kong Industrial AI & Robotics Centre (FLAIR), where I build transformer- and generative-AI-driven inspection systems and lead multimodal medical imaging research.

Outside the lab, I design and ship AI products end-to-end: a video-surveillance platform, an LLM-driven health app, and a real-time computer-vision system for boxing analytics.

My path runs through autonomous-vehicle perception at Dutch Autonomous Mobility, federated-learning healthcare research, and technical writing on generative AI systems — different rooms, same instinct: define the problem precisely, then build the smallest thing that actually solves it.

◆ For the lab

  • Explainable, privacy-preserving model design
  • Federated learning & multimodal fusion
  • Vision transformers & generative-AI research

◆ For the team

  • Production computer vision at real-time frame rates
  • Systems engineered for real sensors & real latency budgets
  • LLM-agent orchestration in production
Research

Published work, funded platforms.

Peer-reviewed methodology and grant-funded research spanning industrial inspection and healthcare AI.

Computer Vision Researcher

Hong Kong Industrial AI & Robotics Centre (FLAIR) · Hong Kong (Remote) · 2024 – Present
  • Architected transformer-based feature extraction pipelines fused with generative AI models for geometric and surface defect inspection, improving defect localization accuracy under real-time, edge-deployment constraints.
  • Built modular, explainable inference workflows using LangChain, orchestrating distributed AI agents across multi-stage industrial inspection subtasks.
  • Led multimodal medical imaging research fusing clinical metadata with visual features, improving early-detection accuracy for high-throughput diagnostic screening, in collaboration with Dr. Irshad Ibrahim.
FUNDED · HK POLYU

RevoCare — AI-Powered Predictive Health Platform, Collaborated with Dr. Irshad Ibrahim

Problem

Chronic-disease patients need continuous, personalized monitoring without compromising data privacy.

Solution

Federated learning-based healthcare analytics using non-invasive biosignals, with privacy-preserving pipelines and LangChain-based LLM clinical guidance.

Impact

Predictive recovery-analytics models enabling intelligent intervention recommendations and reduced readmission risk.

SUPPORTED · HKSTP

InnoGuard AI — Industrial Safety & Monitoring, Collaborated with Dr. Irshad Ibrahim

Problem

Hazardous industrial environments require real-time, interpretable safety monitoring.

Solution

Thermal imaging, vision transformers, and temporal anomaly detection, with generative AI modules producing interpretable anomaly descriptions.

Impact

Enhanced situational awareness and regulatory compliance via proactive, predictive-maintenance-driven interventions.

Farooq, U. et al. "A Hybrid Deep Learning and Large Language Model Architecture for Automated Medical Image Segmentation: From Detection to Pre-Diagnosis."

IEEE Transactions on Artificial Intelligence, 2026
UNDER REVIEW

Farooq, U. et al. "RT-DefectNet: An Intelligent Real-Time Framework for Surface Anomaly Detection in Industrial Systems"

IEEE Access, 2026
UNDER REVIEW
Experience

Where the systems had to survive contact with reality.

Technical Writer

Technolynx · Hungary (Remote) · 2023 – 2024
  • Authored technical white papers and tutorials on LLMs, AI agent architectures, and prompt engineering.
  • Translated advanced generative-AI research into structured, actionable documentation for engineering and business audiences.

Computer Vision Engineer

Dutch Autonomous Mobility · Netherlands (Remote) · 2020 – 2023
  • Engineered end-to-end computer vision pipelines for autonomous shuttles, delivering robust real-time perception under real-world driving conditions.
  • Fused LiDAR and camera data for multi-object tracking using Kalman filtering and SLAM-based localization.

Data Scientist

Dutch Autonomous Mobility · Netherlands (Remote) · 2018 – 2019
  • Built predictive analytics models with TensorFlow and scikit-learn for operational optimization and fleet planning.
  • Developed executive-facing visualization dashboards in Tableau and Power BI.

Computer Vision Intern

Giscle Systems Pvt. Ltd. · India (Remote) · 2017 – 2018
  • Developed a CNN-based object detection system optimized for South Asian road environments, achieving 76% accuracy via transfer learning.
  • Preprocessed and managed large-scale visual datasets with Python and OpenCV.
Products — Built & Shipped Independently

AI I built the whole way through.

Independent products taken from problem definition to a running system — no team, no handoff.

ModSense

In Progress
Problem

Manual video monitoring causes alert fatigue and missed incidents from high false-positive rates.

Solution

A computer-vision SaaS platform that distinguishes genuine threats from routine motion in real time, replacing manual review with automated, evidence-backed alerting.

Computer VisionReal-TimeSaaS

GlucoSense

In Progress
Problem

Chronic-condition patients get isolated data snapshots between visits, with no ongoing intelligent interpretation.

Solution

Continuous biosignal tracking paired with an LLM-driven reasoning layer that turns raw readings into personalized, explainable guidance.

LLM ReasoningHealth AIProduct Design

RingScore Vision

In Progress
Problem

Boxing scoring is manual and judge-dependent, with no objective, real-time record of performance.

Solution

A live CV system fusing YOLOv8 punch detection with MediaPipe Pose over a multi-threaded RTSP pipeline, broadcasting scores via WebSocket.

YOLOv8MediaPipeWebSocket
Skills

The toolkit behind both tracks.

Deep Learning

PyTorch · TensorFlow · Keras · Vision Transformers · CNNs · Transfer Learning

Computer Vision

OpenCV · YOLOv8 · MediaPipe · Segmentation · Multi-Object Tracking · SLAM · Sensor Fusion

Generative AI / LLMs

LangChain · Prompt Engineering · RAG · Multi-Agent Orchestration · Clinical Decision Support

MLOps / Deployment

Docker · MLflow · Git · CI/CD · REST APIs · FastAPI · WebSocket · Edge Deployment

Languages

Python · C++ · R · MATLAB/Octave · LaTeX · Bash

Data & Infra

Pandas · NumPy · Tableau · Power BI · Linux · CUDA · NVIDIA GPU Optimization

Education & Certifications

B.S. Computer Science

University of Engineering and Technology, Pakistan · 2015 – 2019
GPA 3.55/4.00 · Summa Cum Laude
  • Machine Learning in Production — DeepLearning.AI
  • Accelerating End-to-End Data Science Workflows — NVIDIA
  • Medical Image Processing — University of Waterloo
  • Deep Learning for Business — Yonsei University
  • Data Scientist Nanodegree — Udacity

Awards

  • Highest Academic Distinction (Summa Cum Laude) — UET, 2019
  • 2nd Place, PASTIC–UAF Project Competition — 2019
  • 2nd Place, Project Exhibition — UET, 2019
  • Merit-Based Scholarship, Data Scientist Nanodegree — Udacity, 2020
  • Prime Minister's Laptop Scheme Award — 2015
Contact

Let's talk.

Open to research collaborations and senior AI / computer vision engineering roles.