AI Engineer

Over two decades of experience in ML/NLP, including finetuning transformer models, designing production-ready LLM and RAG systems, implementing guardrails and safety mechanisms, prompt engineering, and technical leadership across complex AI initiatives

Led the development of a scientific study/learning copilot designed to help researchers and scientific advisors navigate current literature, compare findings, and understand shifts in scientific consensus — including RAG pipelines, evaluation frameworks, guardrails, and observability

Built ethically aligned AI applications to support psychotherapy and Internal Family Systems (IFS)-inspired workflows, prioritizing human agency, safety, and responsible technology use

Conducted transformer-based machine learning research for predicting ADMET properties from SMILES molecular representations, contributing directly to drug-discovery and molecular-behavior modeling

Created a study/learn copilot and support system for scientific advisors, handling the whole process from design to implementation

Used different large language models (LLMs) in multiple steps (agent-like), Retrieval-Augmented Generation (RAG), evaluation, observability, guardrails, tests and regression tests

Implemented careful testing frameworks to guarantee reliable and controlled model outputs, while preserving human agency and augmenting capabilities through AI

Co-created the company’s offerings, defined goals, and built the online presence

Developed AI applications to support Internal Family Systems (IFS) and talk therapy, ensuring ethical guidelines were respected

Designed and implemented NLP/ML-based solutions, including embedding generation, model testing, and data preprocessing

Conducted workshops, client training, and gathered requirements to align AI tools with business needs

Applied prompt engineering to refine AI-driven solutions for specific business challenges

Developed a hands-on curriculum covering programming and AI applications

Mentored students on building AI prototypes and solving real-world problems

Built applications that use large language models (LLMs) and GPT to automate content creation and categorisation

Applied advanced prompt-engineering strategies to streamline automated workflows

Designed NLP systems to automate advanced customer-support workflows, ensuring high accuracy and efficiency

Researched, developed, and compared transformer-based architectures to predict ADMET (absorption, distribution, metabolism, excretion, toxicity) features from SMILES strings

Adapted published research code for production environments

Implemented techniques to extract actionable insights from emails and transcripts

Focused on fine-tuning large language models, prompt engineering, and simpler NLP methods

Built infrastructure on AWS and secured tools to accelerate data-scientist workflows

Packaged models and implemented CI/CD pipelines

Designed micro-services for ML models, built pipelines, and optimised infrastructure for NLP, image-processing, and OCR tasks

Integrated fine-tuning and prompt engineering across NLP pipelines

Additional contributions: Led team-facilitation workshops and unconferences to enhance knowledge sharing; taught advanced workshops on Python, ML pipelines, and applied NLP for companies and individuals

Performed software engineering for translation and localisation of Siri’s user interfaces and back-end into Italian

Worked on data analysis and data mining using personally developed tools in Perl and JavaScript