Rechner für Lohnkosteneinsparungen
Erfahren Sie, wie viel Sie mit den Recruiting-Lösungen von Team Up sparen können
Berechnen →
Erfahren Sie, wie viel Sie mit den Recruiting-Lösungen von Team Up sparen können
Berechnen →
Book this talent
Book this talent
Senior
Turkey
Mohammad Sadr is a Senior AI/ML Engineer with over five years of professional experience, specializing in LLM development, RAG pipelines, and AI infrastructure. Currently leading a small AI/ML team, he combines hands-on technical expertise with architectural design, building scalable, reliable, and secure AI systems. His experience spans enterprise-grade multi-agent orchestration systems, RAG workflows, and end-to-end AI pipelines from data ingestion to model deployment. He is highly proficient in Python (primary) and has working knowledge of Java, with strong experience in vector databases such as Pinecone, Milvus, and PGVector, as well as agent frameworks including LangChain and LangGraph. Mohammad is also skilled in MLOps and infrastructure tools such as Kubernetes, Docker, Helm, MLflow, Prometheus, and Grafana, enabling him to build, monitor, and optimize robust AI platforms. Motivated by complex AI challenges, Mohammad has successfully improved model performance and operational efficiency through iterative development, CI/CD enhancements, and mentoring junior developers. He demonstrates strong problem-solving skills and a structured approach to designing scalable AI systems.
Computer Science
Computer Science
• Built and deployed an end-to-end ML infrastructure using Kubernetes, Docker, and OpenStack Swift, integrating object storage with shared GPU compute clusters (NVIDIA) to support scalable training, inference, and data pipelines.
• Fine-tuned and optimized multiple LLMs for domain-specific applications in finance and petrochemistry using Hugging Face Transformers, PEFT, and LoRA; achieved an average 22% accuracy improvement on specialized NLP tasks.
• Implemented RAG workflows using LangChain, LLMs embeddings, pgvector, and OCR models such as Tesseract, enabling dynamic retrieval, document chunking, and context-aware generation for enterprise search and Q&A systems, supporting real- time search over 1K+ documents with <1s response time.
• Designed and implemented AI Agent systems using LangGraph, Model Context Protocol (MCP), and event-driven orchestration; enabled autonomous task planning, memory management, multi-agent collaboration, and tool-augmented reasoning across distributed environments.
• Led a team of 4 junior engineers; reduced development time by 25% through improved DevOps workflows and structured code reviews. November 2020–August 2024 AI/ML Engineer GreenWeb Co. (IranGPU SBU), Tehran, Iran GreenWeb, through its IranGPU division, provides GPU-accelerated cloud computing services to AI researchers and startups in Iran.
• Designed and managed a scalable AI/ML development platform, integrating industry-standard frameworks such as TensorFlow, PyTorch, and Scikit-learn to accelerate experimentation, prototyping, and deployment of machine learning models.
• Developed automation scripts and Ansible playbooks that cut infrastructure setup time from 3 hours to 20 minutes.
• Engineered and optimized Containerized-based ML infrastructure, utilizing Docker, and Kubernetes to enable reproducible and automated deployment of end-to-end ML pipelines. Improved developer velocity and system reliability across distributed AI workloads.
• Designed and managed a scalable AI/ML development platform, integrating industry-standard frameworks such as TensorFlow, PyTorch, and Scikit-learn to accelerate experimentation, prototyping, and deployment of machine learning models.
• Developed automation scripts and Ansible playbooks that cut infrastructure setup time from 3 hours to 20 minutes.
• Engineered and optimized Containerized-based ML infrastructure, utilizing Docker, and Kubernetes to enable reproducible and automated deployment of end-to-end ML pipelines. Improved developer velocity and system reliability across distributed AI workloads.
• Engineered a visual product search engine that enhanced user experience. Applied computer vision methods with OpenCV and deep learning using MXNet to enable intuitive, image-based product discovery, boosting user engagement by 15% and click- through rates by 20%.
• Built and deployed scalable classification and clustering models using real-world product data collected from Digikala, a major e-commerce platform. Developed a reliable inference API using MXNet and PyTorch to serve models with high accuracy and performance in categorization tasks
Delivered a critical CMS website solution that empowered the company with seamless publishing capabilities for services and articles, improving content dissemination efficiency. Developed the platform using PHP, WordPress, HTML, CSS, and JavaScript, ensuring an engaging and accessible digital presence
Talent Pool
