Open to opportunities

Alessandro Marci

@alessandromarci

Message

AI security and alignment specialist combining adversarial LLM research with program-level project management.

United Kingdom

Message

What I'm looking for

I seek roles where I can lead adversarial LLM research and operationalise safety findings into governance-ready deliverables, collaborating with technical and policy teams.

I am an AI security and alignment specialist who conducts independent adversarial research on frontier LLMs and translates findings into actionable governance and safety outputs. I design reproducible evaluation pipelines, taxonomies of attack vectors, and forensic case studies for confidential technical review.

My work has systematised 25+ adversarial vectors, quantified exploit efficiency with token-cost curves and turn-based metrics, and produced stakeholder-ready documentation mapping technical vulnerabilities to policy frameworks such as the UK/EU AI Act and NIST AI RMF.

With prior experience managing quality and AI training programs, I bring proven project management skills—Agile, Jira, Asana—and a track record of maintaining high accuracy, reducing budget deviation, and improving operational processes to help teams harden model safety and deploy robust mitigation strategies.

Experience

Work history, roles, and key accomplishments

Current

AI Security & Alignment Specialist

Current

Independent / Freelance

Jan 2024 - Present (1 year 9 months)

Conducted independent research into novel vulnerabilities in frontier LLMs, systematising 25+ adversarial vectors and developing reproducible forensic evaluation frameworks to translate findings into governance-ready mitigations.

LLM Behaviour Analysis Forensic Evaluation Frameworks Prompt Engineering Risk Scoring Technical Documentation

Quality Manager

Scale (Outlier)

Nov 2022 - Jan 2024 (1 year 2 months)

Managed QA and delivery for large-scale AI refinement projects, maintaining a 98% pre-submission accuracy rate, reducing budget deviations by 5%, and coordinating thousands of contributors across Agile and Waterfall workflows.

Quality Assurance Agile)Jira Asana Project Tracking Metrics Analysis

AI Training Analyst

Scale (Outlier)

Dec 2021 - Nov 2022 (11 months)

Trained and evaluated AI models using RLHF, fine-tuning, and transfer learning, achieving 100% dataset readiness through rigorous preprocessing and automated validation while ensuring regulatory compliance.

Model Fine Tuning data preprocessing Validation SpaCy NLTK

Education

Degrees, certifications, and relevant coursework

Anglia Ruskin University

Bachelor of Science with Honors, International Business Management

Activities and societies: Recipient of Top Achievers Award in Human Resources Practices (3rd Semester, 2020) and Top Achievers Award in Business Strategy (3rd Semester, 2021).

Bachelor of Science with Honors in International Business Management awarded in 2022 with recognition for top academic achievement in HR Practices and Business Strategy.