Atlanta, GA | (404) 433-4290 | zarrabinikou@gmail.com
LinkedIn: nikouzarrabi |
GitHub: nkzarrabi
Applied AI/ML Engineer with MS in Computer Science (Interactive Intelligence) specializing in machine learning model training, image-based computer vision, and production systems. Experience with supervised ML model development on structured and spatial datasets, implementing probability map generation, and deploying inference pipelines via REST APIs. Background in document computer vision (OCR), visual reasoning tasks, and building scalable production software with Python, Django, and cloud platforms.
Applied ML & Computer Vision Engineer (Football
Analytics)
Contractor | 2023 – 2024
- Trained supervised ML models on structured and spatial football event
data (Opta, Wyscout format), performing feature engineering and
parameter tuning for predictive analytics - Built regression models for
player performance metrics and market value estimation, generating
out-of-sample predictions integrated into analytics dashboards -
Reconstructed pixel-aligned probability maps from event-level data to
replicate Expected Goals (xG) visualizations and heatmap analytics -
Validated ML model outputs against image-space ground truth, ensuring
spatial accuracy of probability surface generation - Deployed inference
outputs via REST APIs for integration into downstream analytics
systems
iOS Developer
Swimming Cars | Jan 2025 – Present
- Built production SwiftUI iOS app for carpool coordination serving 50+
users, integrating Firebase for real-time data synchronization and cloud
persistence - Implemented geolocation tracking system using Apple Maps
SDK with event-driven architecture for state management - Developing
REST API integration layer for model inference endpoints and ML service
integration - Deployed cloud-native mobile application with zero crashes
through comprehensive testing and CI/CD automation
DevOps Engineer / E-commerce Tech Lead
Keivan Woven Arts | May 2024 – Jan 2025
- Built and deployed automated infrastructure for product listings
across 3 platforms (Shopify, Chairish, 1stdibs), reducing manual
processing time by 80% - Implemented CI/CD pipelines using GitHub
Actions with automated testing and validation, reducing deployment
errors by 95% - Integrated AWS S3 with Django backend for scalable image
storage serving 1,000+ SKUs, improving performance by 60% - Developed
Python automation scripts for cross-platform inventory synchronization
and data validation
Data Automation Intern
Porsche Cars North America | Jun 2018 – Nov 2018
- Developed OCR-based document processing pipeline for invoice
automation, extracting structured data from scanned invoices with 99%
accuracy across 500+ documents per week - Implemented PDF parsing system
using OCR technology to digitize paper-based workflows, reducing manual
processing time by 75% - Built production computer vision pipeline
integrating OCR outputs into automated workflows, eliminating 10,000+
pages annually during critical business transition
Software Engineer (Full Stack &
E-commerce)
Keivan Woven Arts | Jun 2021 – Dec 2022
- Developed Django-based product catalog system with REST APIs managing
1,000+ SKUs across multiple sales channels - Built Python automation
scripts using Playwright for web scraping and testing, reducing QA time
by 70% - Implemented data validation pipelines and error handling
systems, reducing product listing errors by 85% - Designed and deployed
RESTful APIs for platform integration (Shopify, Chairish, 1stdibs) with
JSON data serialization
AI-Powered Employment Assessment Platform | GitHub
- Trained supervised ML models for candidate assessment using
TensorFlow, performing feature engineering and model tuning for
role-specific evaluations - Developed predictive scoring system with
bias-free evaluation criteria, integrating trained models into
production inference pipeline - Built React frontend and Django backend
REST API for serving model predictions and supporting real-time
assessments
Visual Reasoning AI System | Georgia Tech
Coursework
- Implemented visual reasoning algorithms solving Raven’s Progressive
Matrices using PIL for image processing - Developed pattern recognition
and spatial relationship extraction directly from image data -
Demonstrated symbolic AI approach to visual reasoning, identifying
geometric transformations and logical patterns without deep learning
Django Health Tracker (Raspberry Pi) | GitHub
- Full-stack health monitoring application with Django backend and
JavaScript frontend, deployed on Raspberry Pi - Implemented real-time
data visualization and threshold-based alerting for personal health
metrics tracking
Package Sorting Algorithm | GitHub
- Developed classification logic for robotic automation system with
comprehensive test coverage - Implemented rule-based algorithm for
package categorization based on dimensional and weight constraints
MS in Computer Science — Interactive
Intelligence
Georgia Institute of Technology, August 2023
Coursework: Artificial Intelligence, Knowledge-Based AI (Visual
Reasoning), Machine Learning, Robotics: AI Techniques (Perception,
Spatial Reasoning), Software Architecture & Analysis
BS in Mathematics
University of Georgia, May 2018