Machine Learning & Multimodal AI Engineer

Omid Rajabi Rostami

Virginia Beach, VA, USA **********.****@*****.*** 757-***-****

LinkedIn GitHub

PROFESSIONAL SUMMARY

Machine Learning Engineer with expertise in computer vision, multimodal AI, Large Language Models (LLMs), and trustworthy artificial intelligence. Experienced designing, developing, optimizing, and deploying production-oriented machine learning systems using PyTorch, Hugging Face Transformers, ONNX Runtime, and cloud platforms. Proven experience building end-to-end AI solutions for document intelligence, OCR, intelligent surveillance, remote sensing, and geospatial analytics. Strong software engineering background with a passion for delivering scalable, high-performance AI applications. TECHNICAL SKILLS

Programming Languages: Python • Java • C# • SQL • JavaScript Machine Learning: PyTorch • TensorFlow • Hugging Face Transformers • CNNs • Vision Transformers • Computer Vision • Image Segmentation • Object Detection • OCR • Distributed Training • Model Optimization • Deep Learning Generative & Multimodal AI: LLaVA • Mistral • CLIP • Retrieval-Augmented Generation (RAG) • Prompt Engineering • Multimodal AI Deployment & MLOps: ONNX Runtime • Google Cloud Platform • Model Deployment • Unit Testing • Software Quality Software Engineering: REST APIs • Git • Linux • Object-Oriented Programming • Agile Development Data: NumPy • Pandas • OpenCV • ArcGIS Pro • Matplotlib PROFESSIONAL EXPERIENCE

Machine Learning Engineer / Research

Old Dominion University Research Foundation (ODURF) Norfolk, VA Aug 2021 – September 2026

• Engineered multimodal AI systems integrating OCR, CLIP, LLaVA, and Mistral-7B to improve document understanding and OCR correction for degraded documents.

• Built end-to-end deep learning pipelines using PyTorch for adversarially robust image segmentation and remote sensing applications.

• Improved OCR accuracy by 8% through image enhancement and multimodal reasoning techniques.

• Automated machine learning workflows, experiment tracking, data preprocessing, and model evaluation.

• Optimized inference performance using ONNX Runtime for accelerated deployment.

• Applied distributed training and model optimization techniques to improve scalability and GPU utilization.

• Developed FastAPI inference services for machine learning models and containerized applications using Docker.

• Utilized Hugging Face Transformers to build and evaluate transformer-based vision-language models.

• Maintained software quality through Git, unit testing, and reproducible development workflows. AI Software Engineering Intern

IPConfigure Norfolk, VA Jan 2026 – Mar 2026

• Developed computer vision algorithms for intelligent surveillance and real-time video analytics.

• Built image-processing pipelines using PyTorch for production surveillance applications.

• Optimized model inference for live video streams using ONNX Runtime.

• Collaborated with software engineers to integrate machine learning models into production software.

• Improved runtime performance through model optimization and efficient inference pipelines. Machine Learning Engineer Intern

City of Norfolk Norfolk, VA Jun 2022 – Sep 2022, Jun 2023 – Sep 2023

• Developed deep learning models for flood prediction using satellite imagery and ArcGIS data, achieving 90%+ prediction accuracy.

• Built scalable machine learning pipelines for geospatial data processing and feature engineering.

• Designed REST APIs and deployed machine learning services using Spring framework.

• Created Python dashboards supporting data-driven decision-making for city planning.

• Prototyped natural-language geospatial search using lightweight LLM prompting. Instructor (Part-time)

University of Hormozgan, Bandar Abbas, Iran Aug 2017 – Aug2020

• Assisted in delivering lectures for courses on Python programming, machine learning, and neural networks.

• Provided one-on-one support for students and led group discussions to clarify complex topics.

• Recognized for exceptional student support, earning praise for my contributions to developing the Computational Intelligence course. Full-Stack Software Engineer

Negin Tejarat, Esfahan, Iran Jan 2014 – Jan 2017

• Designed and implemented scalable backend systems using Python, SQL, and RESTful APIs.

• Developed enterprise web applications using JavaScript, HTML, CSS, and relational databases.

• Built secure and maintainable software following modern software engineering principles.

• Collaborated with cross-functional teams to deliver production software solutions. SELECTED PROJECTS

Multimodal Document Intelligence Platform

PyTorch • Hugging Face • LLaVA • CLIP • Mistral-7B • FastAPI • Docker

• Designed a multimodal AI pipeline combining OCR, vision-language models, and large language models for intelligent document understanding.

• Improved OCR correction by 8% on degraded and noisy documents.

• Built scalable inference services supporting efficient deployment and evaluation. Secure AI for Remote Sensing

PyTorch • Detectron2 • Transformers

• Developed adversarial attack and defense algorithms for semantic segmentation models operating on aerial imagery.

• Built robust deep learning pipelines to evaluate security and robustness of remote sensing AI systems.

• Conducted large-scale experimentation on transformer-based vision architectures. AI Flood Prediction System

PyTorch • ArcGIS • FastAPI • Docker

• Built a deep learning platform achieving 90%+ prediction accuracy for flood forecasting.

• Integrated geospatial data processing with production-ready REST APIs.

• Developed visualization dashboards supporting real-time environmental monitoring. EDUCATION

Ph.D. in Computer Engineering, Graduated, 2021-2026 Old Dominion University

M.S. in Artificial Intelligence, Graduated, 2017-2019 University of Hormozgan

B.S. in Software Engineering, Graduated, 2010-2015 Payam Noor University

PUBLICATIONS

Google Scholar: https://scholar.google.com/citations?user=D01-Wn4AAAAJ&hl=en

Contact this candidate