Abdellaoui Houssem

** AI/NLP Product Owner Own end-to-end delivery of Plainly, a plain-language writing assistant for regulated financial/non-financial communications “1st solution of scoring + plain-language writing support”. Drive roadmap, backlog prioritization, and KPIs; translate user needs into user stories/acceptance criteria across Engineering, DS, Legal/Compliance, and Business. Deliver agentic GenAI products with guardrails: CSRD Drafter (structured workflow) → 30% reduction in time-to-first-draft for ESG experts. Ship & improve auditable RAG pipelines: TAW RAG → 75% reduction in transparency-criteria collection time; monitor quality in production. ** AI Governance Authored and rolled out a company-wide AI Code of Conduct (tools, HITL verification, PII/privacy, IP/copyright, disclosure). Chaired cross-functional governance (Legal/Compliance, Security, Product, Engineering) and implemented exceptions + incident/violation reporting + annual review cycle. Trained employees/freelancers on safe GenAI usage (prompting, Copilot best practices); policy adopted by teams across FR/US/PH. Strengthened responsible production GenAI practices (traceability, auditability, guardrails, HITL review).

Led AI/NLP R&D initiatives for B2B services (text analysis, classification, information retrieval). Converted business needs into research roadmaps; took initiatives from exploration to validated POCs/prototypes. Coordinated multidisciplinary squads and academic partners; set milestones, reviews, and stakeholder rituals. Designed evaluation frameworks and prepared handover packages for product industrialization/ integration.

Conducted applied research on Arabic sentiment analysis (dialectal/noisy social media text). Built and evaluated deep learning models for classification and opinion extraction. Supervised student AI/NLP projects (scoping, methodology, implementation, results). Taught computer science/AI courses with hands-on project guidance.