Meriem Dekali

Designed and built a tender-intelligence SaaS that centralises open tenders and automates the analysis of French-language procurement dossiers (CDC/CCAP/CCTP) for go/no-go bidding decisions.

Architected the platform end to end: ingestion/scraping pipeline, RAG document analysis, an ML win-probability scoring model trained on historical bids, and a real-time dashboard (open tenders, active bids, pipeline KPIs, deadline alerts).

Built the stack with Python, FastAPI, LangGraph, PostgreSQL + pgvector, Supabase (auth/realtime/storage) and Celery/Redis workers, using local embedding models to keep inference costs near zero.

Secured and structured the first enterprise client engagement, producing a full bilingual (FR/EN) package — proof-of-concept, solution architecture, implementation plan, legal framework and a comprehensive NDA — spanning product, engineering and commercial delivery.

Business needs framing to production deployment, turning operational challenges into solutions integrated into the company's internal processes.

Design and deploy custom AI/ML models in Python (TensorFlow, scikit-learn) to improve operational efficiency.

Built hybrid, context-aware RAG systems to make internal documents and knowledge searchable and actionable.

Developed sentiment analysis and automated document processing pipelines using Computer Vision and OCR.

Manage business-oriented projects across their full lifecycle, coordinating technical teams and business units from specification to delivery.

Act as the bridge between technical and business stakeholders, translating data outputs into concrete, actionable recommendations.