Youssef Aitousarrah

About Youssef Aitousarrah

I design and deploy production-grade Generative AI systems that transform complex, multi-format enterprise data into instant, reliable answers and scalable marketing content. At C3 AI, I architected multimodal RAG pipelines, integrating query expansion, hybrid dense/BM25 search, reranking, and long-context LLMs, slashing knowledge-work cycles from weeks to hours and boosting accuracy from 49% to 85% on customer dataset.As Product Lead for Generative Marketing, I delivered an end-to-end platform for bulk email and content generation, leveraging customer documents, web data, multi-agent workflows, and a validation/feedback loop for continuous improvement. I’m passionate about driving measurable impact, building robust evaluation frameworks, and creating reusable AI playbooks.

Experience

1. Tech Lead - Generative Ai

   C3 AI

   Jan 2025 — Present · Redwood City, CA, US

   Generative AI DS Lead:• Developed and deployed generative AI pipelines for automated information retrieval, troubleshooting, and document generation, reducing turnaround time from weeks to hours and minimizing manual errors.• Architected a multimodal RAG stack with query expansion, hybrid search, reranking, and multimodal answering, productionized with monitoring and evaluation systems.• Improved accuracy from 49% to 85% on a customer dataset by using the new RAG pipeline.• Built a distributed metadata extraction pipeline to enrich vector-store passages, enhancing search performance and recall.• Enhanced unstructured file parsing and chunking (cross-passage context, bbox margins, neighbor-text merging, robust table parsing) to improve passage quality and answer faithfulness.• Experimented with long-context models to increase accuracy while reducing similarity-search latency.• Implemented hypothetical query expansion to broaden recall without compromising precision.Generative Marketing DS Lead:• Implemented an end-to-end solution for generating bulk emails and marketing content (Ads, Landing Site, Visuals) using customer documents and web data.• Designed multi-agent workflows to support content generation, competitive analysis, and integrations with MCP tools (Salesforce, Slack,…)• Developed a content validation pipeline with a feedback loop, enabling users to mass-correct outputs and improve quality iteratively.• Successfully delivered multiple enterprise-level Generative AI solutions to Fortune 500 companies, slashing time spent on complex tasks and unlocking substantial economic value.https://c3.ai/blog/solving-the-engineering-modeling-and-simulation-bottleneck-in-building-design-with-agentic-ai/?utmMedium=q216aw & utmSource=8li1o0

Education

• Lycée Ibn Timiya

  Math, Physics, Computer sciences and Engineering Science

  2014 — 2016

• Stanford University

  Master of Science - MS, Energy Resources Engineering (ERE)

  2019 — 2021

• École Polytechnique

  Math, Economics, Physics, Mechanics and Computer Science

  2016 — 2020