Muhammad Shalaby

Led research, architecture, and deployment of multiple strategic AI agents to enhance business development and operational efficiency. Architected end-to-end AI agent pipelines: data sourcing (HubSpot, blogs, Excel), transformation (Python, Polars), leveraging LLMs (OpenAI, Gemini), RAG systems, and web search (Tavily, Jina AI). Optimized agent performance through model selection and data augmentation; reduced token consumption with multi-optimization techniques. Built a production Slack bot with custom prompt chains, knowledge base, and autonomous workflows (Python, Dify). Documented architectures and processes to enable effective knowledge transfer.

Led research, architecture, and deployment of multiple strategic AI agents enhancing business development and operational efficiency. Architected end-to-end AI pipelines: data sourcing (HubSpot, blogs, Excel), transformation (Python, Polars), leveraging LLMs (OpenAI, Gemini), RAG systems, and web search (Tavily, Jina AI). Optimized model selection, prompt engineering, and data augmentation to improve agent precision and reduce costs. Built production-ready Slack bot with custom prompt chains, knowledge base, and autonomous workflows; documented architectures to enable knowledge transfer.

Revitalized a struggling pharma QA platform plagued by technical debt and poor performance. Led GraphQL adoption (workshops, guides, implementation); orchestrated migration to Syncfusion UI; implemented offline data sync. Restored stakeholder confidence, improved load times, and reduced memory usage and CI/CD effort, enabling faster feature delivery.