I am Reza Zeraat, an AI Engineer specializing in generative AI, machine learning, and full-stack development. I bring hands-on experience with Python, React, JavaScript, C#, and frameworks such as TensorFlow and PyTorch. My portfolio emphasizes both the technical depth required for AI/ML model development and the collaborative mindset needed for end-to-end solution delivery. I actively seek opportunities to work on challenging projects—ranging from data pipelines to agentic AI systems—and to share knowledge with the broader AI community through blogs, open-source contributions, and technical discussions.

I am a passionate AI Engineer focused on designing and deploying scalable AI/ML solutions that drive real-world impact. My technical toolkit includes Python for deep learning (TensorFlow, PyTorch), JavaScript/TypeScript for frontend development (React), and C# for backend services. I thrive on solving complex problems in areas like computer vision, natural language processing, and generative models. I hold a strong foundation in data preprocessing, model optimization, and deployment best practices (e.g., Docker, Kubernetes). My interests span full-stack development, data engineering, and automated ML pipelines, which I combine to build production-grade AI applications.

I am currently deepening my expertise in generative AI, specifically focusing on fine-tuning large language models (LLMs) and exploring prompt engineering techniques. I regularly experiment with Retrieval-Augmented Generation (RAG) to build context-aware applications, such as AI-driven chatbots and recommendation engines. I have completed courses on modern AI frameworks and continuously refine my skills by contributing to open-source AI repositories and participating in Kaggle competitions.

• Machine Learning & Deep Learning: TensorFlow, PyTorch, Scikit-Learn, Keras.
• Generative AI & NLP: Transformer architectures, prompt engineering, fine-tuning LLMs.
• Computer Vision: Convolutional Neural Networks (CNNs), object detection, image segmentation.
• Data Engineering & MLOps: Data pipelines (Apache Airflow), Docker, Kubernetes, CI/CD for ML.
• Full-Stack Development: React, Node.js, Express.js, C#, RESTful APIs, GraphQL.
• Tools & Platforms: Git/GitHub, AWS SageMaker, Azure Machine Learning, Google Cloud AI Platform.

Sharing knowledge accelerates innovation. I write technical blog posts detailing my approaches to AI model optimization, RAG implementations, and prompt engineering. I actively contribute to AI-focused GitHub repositories, submit pull requests to open-source ML libraries, and engage in code reviews to uphold best practices. I have participated in AI hackathons and community events, where I collaborate with multidisciplinary teams to prototype AI-driven solutions under tight deadlines.

• LinkedIn: Reza Zeraat ([GitHub][7], [GitHub][8])
• Email: [email protected] ([GitHub][7], [GitHub][8])
• GitHub: @rezacr588 ([GitHub][7], [GitHub][8])

Feel free to reach out to me if you have AI/ML project ideas, open-source collaboration opportunities, or want to discuss the latest trends in generative AI.

Below are highlights of some AI-centric projects showcasing my problem-solving skills and end-to-end system development:

1. Generative Chat Agent (LangChain & RAG):

   • Implemented an AI agent using LangChain to handle natural language queries and integrate real-time data sources.
   • Employed vector databases (Weaviate) for embedding storage, enabling efficient retrieval of context for user interactions.
   • Deployed the solution as a Docker container orchestrated with Kubernetes, achieving horizontal scalability.

2. Computer Vision Pipeline for Anomaly Detection:

   • Designed and trained a CNN-based model using PyTorch to detect anomalies in manufacturing imagery.
   • Automated data ingestion with Apache Airflow, ensuring continuous model retraining on new labelled data.
   • Integrated a RESTful API backend (Node.js) and React frontend to visualize real-time detection results.

3. NLP Sentiment Analysis Dashboard:

   • Built an end-to-end pipeline to collect social media data, preprocess text (tokenization, lemmatization), and train transformer-based sentiment classifiers.
   • Developed a React dashboard with D3.js visualizations for real-time sentiment tracking across keywords and topics.
   • Containerized the service with Docker and deployed it on AWS Fargate for high availability.

4. Reinforcement Learning for Game AI:

   • Implemented a DQN agent in TensorFlow for an OpenAI Gym environment to optimize strategy in a turn-based game.
   • Biases, achieving a 20% performance improvement over baseline.

I welcome collaboration on AI-driven initiatives—whether you have a novel startup idea, need a machine learning consultant for an enterprise project, or want to co-author a technical blog. Drop me a message on LinkedIn or open an issue in any of my GitHub repositories to start a conversation about how we can create intelligent, impactful solutions together.