Rami Mashkouk

- Engineered the core backend architecture for "Chatsky-UI" (an AI-Agent builder) using FastAPI and Docker. Delivered a production-ready, scalable backend that enabled the product’s commercialization and secured a long-term customization contract with a major enterprise client (financial sector).

- Orchestrated CI/CD pipelines and release cycles using GitHub Actions, reducing deployment friction and ensuring seamless integration of the backend with external microservices.

- Established engineering standards for the product team by enforcing rigorous code reviews and merge management, which improved code stability and minimized regression issues during rapid development phases.

- Architected "DeepPavlov-KG", a custom Python library for abstracting Knowledge Graph operations. The library was adopted across the Dream platform ecosystem, reducing backend integration time by 45% for new data sources.
- Enhanced the "Dream" chatbot platform by implementing an LLM-powered memory system. Designed the API to read/write context dynamically, significantly improving user engagement and conversational continuity.

Implemented and evaluated transformer-based models from Hugging Face for advanced entity extraction and relationship modeling, comparing their performance with traditional NER approaches.

- Contributed to the engineering of a logistics backend that applied a genetic algorithm to solve complex task distribution problems, collaborating with business analysts to align solutions with real-world operational needs.
- Enhanced productivity by deploying an automated daily harvesting scheduling system that optimized operational efficiency across departments.