Moaed Yahia

Spearheaded the modernization of Lebara's global eCommerce platform, transforming a legacy architecture into a high-performance, scalable system using React, TypeScript, Next.js, and GraphQL. This migration improved page load speed, stability, and developer productivity across international markets serving millions of users.

I designed the front-end architecture around SSR and Incremental Static Regeneration (ISR), boosting PageSpeed Insights from 30% to over 90%. I also built a reusable component library and a fully versioned design system in collaboration with UX teams, utilizing Storybook and Figma, which reduced UI development time and improved cross-team consistency.

On the back end, I architected Node.js microservices powered by PostgreSQL and MongoDB, improving API reliability and enabling faster feature delivery. I introduced modern DevOps practices, including Docker, Jest, CI/CD, and ESLint/Prettier standards, strengthening quality and reducing defects across the engineering organization.

This migration became the foundation for the company's new digital experience and significantly improved performance, maintainability, and scalability.

Headed the end-to-end development of a multilingual AI voice assistant and RAG chatbot deployed across a major European telecom platform serving 5+ million users. The solution combined Twilio Voice, Deepgram STT, ElevenLabs TTS, and OpenAI models to create a speech-to-speech workflow capable of verifying customer identity in six languages with 90% first-call resolution.

I architected the system's LLM pipeline, built memory-augmented RAG retrieval with Azure Cognitive Search, and implemented multi-turn context retention for more human-like interactions. The chatbot increased digital self-service containment from 35% to 72% within three months.

I collaborated with data, QA, and compliance teams to introduce automated regression suites, model drift monitoring, and high-availability deployment patterns, ensuring zero critical production defects and reducing the contact center load by 50%.

Designed and delivered a large-scale AI-powered customer support chatbot for Lebara, serving over 5 million users across multiple regions. As AI engineering lead, I led the development of a retrieval-augmented generation (RAG) chatbot integrated with Azure Cognitive Search, embeddings, and conversation memory to provide accurate, context-aware responses across customer channels.

The chatbot was deployed across web and support platforms, enabling customers to resolve account, billing, and service-related queries autonomously. I also implemented automated data pipelines for embedding generation and indexing, ensuring the knowledge base stayed continuously up to date. The solution significantly improved self-service capabilities, increasing containment rates from 35% to 72% within three months and reducing dependency on live agents.

Additionally, I introduced governance frameworks, monitoring, and regression testing to maintain high-quality responses and system reliability at scale.