About Me
I am Tadhg Papillaud Looram, a hands-on AI Engineer and Technical Leader who specializes in the engineering of machine learning systems that are effective, measurable, and reliable in practice.
I’m the Co-Founder and CTO at Portex, where I lead the AI vision and execution across evaluation systems, observability, and agentic workflows. I built Asymmetry Zero, an open-source evaluation framework for benchmarking LLMs and agents on domain-specific tasks, and helped lead the development of COMPOSITE-STEM, a frontier benchmark of expert-authored scientific tasks built on Portex’s Datalab and released with tooling for agent evaluation.
I have a background in Machine Learning and Language Models, Evaluation Systems, and Deployment. My strength is rapid comprehension of requirements, prototyping of research and emerging toolchains, and end-to-end ownership of the entire process in a fast-paced setting.
Prior to founding Portex, I earned my Master’s Degree in Data Science at Harvard University, conducting NLU application research in relation to bytecode analysis and working as a Graduate Teaching Fellow for graduate level courses.
At the Federal Reserve Bank of New York, I spent five years in the Markets Group applying machine learning and data-driven methods to risk management problems across portfolio surveillance, transaction monitoring, and payments risk. My work included developing models and analytical workflows for anomaly detection, AML, sanctions screening, and beneficiary identification in high-stakes operational settings.
From industry and academia through to startups, my personal passions are rooted in curiosity and a bias towards action. The thing that makes me most excited about the space is being close to cutting-edge developments in AI and rapidly implementing them.
Technical Interests:
Agent-based AI
Evaluating and Benchmarking LLMs
Applied LLM Systems
Deployment, Inference, and Observability
Post-Training Workflows
CV available upon request. Please reach out via email at tadhg + . + looram + @ + gmail.com
I am Tadhg Papillaud Looram, a hands-on AI Engineer and Technical Leader who specializes in the engineering of machine learning systems that are effective, measurable, and reliable in practice.
I’m the Co-Founder and CTO at Portex, where I lead the AI vision and execution across evaluation systems, observability, and agentic workflows. I built Asymmetry Zero, an open-source evaluation framework for benchmarking LLMs and agents on domain-specific tasks, and helped lead the development of COMPOSITE-STEM, a frontier benchmark of expert-authored scientific tasks built on Portex’s Datalab and released with tooling for agent evaluation.
I have a background in Machine Learning and Language Models, Evaluation Systems, and Deployment. My strength is rapid comprehension of requirements, prototyping of research and emerging toolchains, and end-to-end ownership of the entire process in a fast-paced setting.
Prior to founding Portex, I earned my Master’s Degree in Data Science at Harvard University, conducting NLU application research in relation to bytecode analysis and working as a Graduate Teaching Fellow for graduate level courses.
At the Federal Reserve Bank of New York, I spent five years in the Markets Group applying machine learning and data-driven methods to risk management problems across portfolio surveillance, transaction monitoring, and payments risk. My work included developing models and analytical workflows for anomaly detection, AML, sanctions screening, and beneficiary identification in high-stakes operational settings.
From industry and academia through to startups, my personal passions are rooted in curiosity and a bias towards action. The thing that makes me most excited about the space is being close to cutting-edge developments in AI and rapidly implementing them.
Technical Interests:
Agent-based AI
Evaluating and Benchmarking LLMs
Applied LLM Systems
Deployment, Inference, and Observability
Post-Training Workflows
CV available upon request. Please reach out via email at tadhg + . + looram + @ + gmail.com