Data Scientist / AI Systems Engineer / Pakistan / Remote

Muhammad Muneeb Arshad

I build computer vision systems and AI automation workflows with production-minded architecture.

work github email

systems

Computer Vision Systems

Model pipelines for industrial inspection, sports analytics, medical imaging, OCR, segmentation, and deployment workflows.

  • - YOLO player/object detection, pose estimation, court keypoints, and ball tracking
  • - OCR and defect workflows with annotation guidelines, preprocessing, and evaluation
  • - Medical image segmentation/classification experiments with clinical-style validation
YOLO DETR CNN U-Net OpenCV PyTorch

AI Automation Systems

Deterministic AI backends and workflow systems with state, validation, human review, observability, and production integration points.

  • - LangGraph intent routing, planner/executor/validator pipelines, and trace replay
  • - Pydantic schema validation, bounded retries, confidence gates, and hard-fail behavior
  • - FastAPI, Redis, Celery, WebSocket dashboards, n8n workflows, and webhook integrations
LangGraph FastAPI Pydantic Redis Celery n8n

selected work

AI Automation Planner / Executor / Validator trace

AI Business Automation Platform

A production-oriented operations platform that classifies, plans, executes, validates, and streams plain-English business tasks to a live dashboard.

Built intent routing, IR generation/validation, schema resolution, planner/executor/validator orchestration, WebSocket progress, and trace replay.

LangGraph FastAPI React WebSocket Redis
repo
Agentic Systems Planner / Executor / Validator loop

AI Developer Agent

A multi-agent developer-ops system that turns natural-language operational requests into planned, executed, validated, and logged workflows.

Built planner, executor, and validator agents with Redis state, Pydantic schemas, bounded retries, hard-stop failure behavior, and FastAPI run/status endpoints.

Python LangGraph FastAPI Redis Pydantic
Computer Vision Vision pipeline map

Padel Analytics System

A training-first sports analytics pipeline covering player detection, pose estimation, ball tracking, and court keypoint detection.

Normalized datasets, converted annotations, trained models on RunPod, adapted inference logic, and prepared CVAT annotation handoff.

YOLO Pose Estimation RunPod CVAT Python
repo
Computer Vision Evidence scoring panel

AI Architecture Plan Auditor

A multimodal audit system that evaluates architectural drawings with vision extraction, structured rules, and evidence-grounded scoring.

Combined GPT-4o Vision, JSON-mode scoring, Markdown rulesets, disk cache persistence, and Streamlit reporting.

GPT-4o Vision Streamlit Python JSON Mode Rulesets
repo
Decision Support Scenario planning workspace

Adaptive AI Office Space Planner

An AI-assisted office planning tool for importing floor plans, editing layouts, scoring scenarios, and comparing workspace decisions.

Built DXF import, SVG-based editing, environmental scoring, comfort diagnosis, scenario comparison, and FastAPI-backed planning flows.

TypeScript React Vite SVG FastAPI
repo

more projects

AI Product

AI Voice Notes to Content

repo

Voice-note-to-content workflow for turning captured ideas into structured drafts and reusable content assets.

HTMLAI WorkflowContent Automation

Automation

n8n Workflow Showcase

repo

Credential-light n8n workflow demos for AI lead triage, sales follow-up, and AI-powered social clip planning.

n8nWorkflow DesignHuman Review

Medical Vision

X-ray Spine Segmentation

repo

Medical imaging pipeline adapting TotalSpineSeg-style segmentation from 3D MRI workflows toward 2D X-ray vertebral analysis.

nnU-NetMedical ImagingPythonSegmentation

LLM Systems

Deterministic AI Processor

repo

Local-first structured extraction system using schema validation, bounded retries, and hard-stop LLM behavior.

OllamaPydanticPythonSchemas

NLP

RAG Document QA System

repo

Context-grounded document question answering with retrieval, chunking, Flask API, and responsive chat UI.

RAGFlaskTF-IDFNLP

Analytics

Retail Sales Data Analytics

repo

Retail sales and profitability analytics dashboard covering pricing, discounts, returns, and product performance.

Power BIDAXPower QueryBI

Business Intelligence

Financial Performance Dashboard

repo

Power BI financial dashboard for revenue, profitability, regional performance, and discount impact analysis.

Power BIDAXKPI Design

Data Engineering

OpenSky Flight Data Analysis

repo

Python pipeline for real-time aircraft telemetry, geospatial EDA, Folium maps, and aviation trend analysis.

PythonPandasFoliumAPI

experience

Dec 2025 - Present

Elunic GmbH

Data Scientist

Industrial Computer Vision & Production ML Systems

  • - Design and deliver computer vision solutions for industrial quality control, from dataset strategy through deployment and monitoring.
  • - Create annotation guidelines, labeling workflows, QA checks, and model training pipelines for robust defect detection and OCR systems.

2025 - Present

Upwork

Freelance AI Engineer

Production AI Systems Across Vision, Automation, and Decision Support

  • - Build AI automation systems with LangGraph, FastAPI, schema validation, human review, Redis/Celery execution, and trace visibility.
  • - Deliver computer vision systems for sports analytics, architecture plan auditing, medical imaging, and manufacturing workflows.

Mar 2022 - Nov 2022

Emumba (Pvt) Ltd

Software Engineer

Full-Stack Development & DevOps

  • - Built production features and responsive React/TypeScript interfaces with maintainable frontend architecture.
  • - Implemented CI/CD pipelines using GitHub Actions to improve deployment reliability and iteration speed.

2020 - 2025

Floreat Enterprises

Managing Partner

Business Operations & Strategic Management

  • - Directed operations, sales strategy, inventory management, client communication, and export delivery for a surgical instruments business.
  • - Introduced structured sales tracking, inventory analytics, and cost analysis frameworks to improve financial visibility.

May 2019 - Jul 2019

TUKL-NUST R&D Centre

Machine Learning Research Assistant

Computer Vision Research & Prototyping

  • - Researched machine learning and computer vision methods for medical image analysis, CNN architectures, and transfer learning.
  • - Developed a prototype for automated skin cancer classification using CNNs and public dermatology datasets.

education

Bachelor of Science in Computer Science

National University of Sciences and Technology (NUST)

2016 - 2020

Final Year Project: Skin Disease Identification using Convolutional Neural Networks

certifications

Machine Learning Specialization
MERN Stack Full Ecommerce Site - React, Redux, Node.js
Front-End Web UI Frameworks and Tools: Bootstrap 4
Microsoft Office Specialist Master Certification