Get to know me!
I'm Ahmed Doghri, a Senior AI Engineer with 8+ years shipping LLM and ML systems that run in production, not just in a demo. I specialize in RAG pipelines, multi-agent orchestration (LangChain / LangGraph), and LLM evaluation harnesses (RAGAS, DeepEval), deployed on production MLOps stacks (Vertex AI, SageMaker, Kubeflow) inside regulated domains — pharma / FDA, banking / fintech, and B2B. Concretely: a Slack RAG assistant that cites its sources at 94% precision, 7 LLM agent pipelines for FDA IND submissions that cut hallucination rates 60%, a multi-agent catalog system that removed 60% of manual tagging, and inference serving 50K+ requests a day at 30% lower cost. As founder of Idea Boundary I own the full arc — discovery, delivery, and measurable business impact — so I'm equally at home in a senior / staff IC seat or an AI-product role. Based in North America (US + Canada work authorization), open to remote or relocation.
The open-source work below is that same discipline applied to current research: small, testable systems where every claim ships with code, a baseline, and a command that reproduces the number — spanning LLM infrastructure and agents through bioinformatics, genomics, vision, and audio.
Professional Timeline:
- Idea Boundary - Founder & Principal AI Engineer (Independent Consulting). Production RAG at 94% citation precision, multi-agent orchestration, and churn prediction for enterprise clients; own the arc from discovery to delivery.
- Weave - Senior AI Engineer. Architected 7 LLM agent pipelines for FDA IND submissions, cutting hallucination rate 60%; shipped table-extraction RAG and cut inference cost 30% at 50K+ requests/day.
- BenchSci - Senior ML Engineer. Transformer + LLM ensembles for biomedical NER on Vertex AI (+30% accuracy), with full MLOps, CI/CD, and DeepEval evaluation.
- IBM - Senior Data Scientist & Consultant. Conversational AI and NLP pipelines for enterprise banking and retail; drove multi-million-dollar contact-center savings.
- Scotiabank - Product Owner. Launched an AI bot for automated negative-news detection, improving risk management and reducing manual monitoring 75%.
- Ford Motor Company - Research Assistant. High-precision misfire detection and DAQ systems for automotive signal processing.
