Hire Paolo G. - Prompt Engineer - San Francisco

Indeed is the world's #1 job platform. Project: Italian-language localization of AI-generated occupational questionnaires using a Translate > Estimate > Refine (TEaR) LLM workflow. - Evaluated and refined LLM prompts across multiple iteration cycles, improving translation output quality for occupational content in Italian. - Normalized thousands of occupational canonical titles by removing redundant data, increasing consistency across the job taxonomy dataset. - Assessed grammar quality and translation accuracy of LLM-generated outputs, flagging errors and working with the ML team to resolve systemic issues. - Contributed to prompt engineering improvements that directly enhanced model performance for Italian-language job content. Tools: GitLab, Atlassian Jira, AWS-S3, Taxonomy Repository, DataLogs

* Amazon Q is Amazon's enterprise conversational AI assistant. * Project: AI model training, dataset evaluation, and annotation quality improvement for Italian-language deep research capabilities. * Led annotation efforts for Amazon Q's deep research AI workflows, generating high-quality prompts and evaluating datasets used to train the model. * Language Linguistic Auditor: Led Italian language quality verification programs for enterprise AI data collection * Advanced AI Model Training: AI model training initiatives for Amazon Q's conversational AI system, coordinating deep research annotation efforts * Data Quality Optimization: Lead data quality optimization initiatives through qualitative error trend analysis, developing and implementing strategic action plans that measurably improved data quality metric * Cross-functional ML Collaboration: Collaborated with ML Data Scientists and Operations Managers to drive process automation innovations * Research-driven Analysis: Lead deep research initiatives in AI/ML annotation methodologies, contributing to enhanced model performance and translation quality standards * Collaborated with ML Data Scientists and Operations Managers to identify bottlenecks and introduce process automation, reducing manual annotation overhead. * Contributed to research on AI/ML annotation methodologies, resulting in enhanced Italian-language model performance and translation quality. * Authored comprehensive linguistic annotation guidelines and best practice frameworks, establishing standardized processes that enabled consistent quality across distributed annotation teams

Meta Reality Labs builds AR/VR hardware and AI-powered products. Project: Italian-language LLM annotation for Meta AI Assistant and Meta Smart Glasses live translation features (ASR, ML, NLP). - Linguistic Auditor: Managed external vendor relationships and quality assurance programs for linguistic annotation services, serving as primary point of contact and conducting comprehensive quality audits with detailed reporting on vendor performance and deliverable - LLM training data annotation/labeling: Coordinated LLM training data annotation and labeling initiatives for META Assistant and META Smart Glasses live translation features, ensuring data quality standards met product requirements across ASR (Automatic Speech Recognition), ML (Machine Learning), and NLP (Natural Language Processing) workstreams - Designed and delivered onboarding, training, and mentoring programs for new team members, establishing structured knowledge transfer processes that accelerated ramp-up time and maintained annotation quality consistency - Standards & Process Documentation: Developed and maintained linguistic annotation guidelines and best practice documentation, creating scalable frameworks that standardized workflows across internal teams and external vendor partners - Leveraged SQL to extract and analyze annotation data for linguistic quality audits - Led speech annotation quality assurance initiatives for external vendor deliverables, establishing audit protocols and quality metrics that ensured compliance with technical specifications and project timelines