Francisco Liaci

FRANCISCO LIACI AI Data & Quality Lead | LLM Training & Evaluation | Multilingual Data Strategy PROFESSIONAL SUMMARY Seasoned AI Data & Quality Lead with 20+ years of expertise spanning linguistics, quality assurance, and structured knowledge systems. Currently directing large-scale human-in-theloop (HITL) data operations critical for LLM training and strategic evaluation. Leverage a deep foundation in translation and linguistic precision to strengthen applied ML workflows, delivering high-signal human data that improves model performance, reliability, and multilingual data integrity. KEY ACHIEVEMENTS • Consistently scored 90%+ in QA evaluations across major global localization platforms. • Improved throughput by 25% via MTPE workflow optimization without compromising quality. • Maintained brand voice and terminology consistency across multilingual campaigns, reaching 98% terminology accuracy in QA reviews. • Partnered cross-functionally to meet client quality/compliance expectations, including ISO 17100-aligned processes. CORE EXPERTISE • • • • • • • • • LLM Training & Evaluation: Strategy, execution, and quality control. HITL Ops & Pipelines: Design, management, and scaling of human data workflows. Applied ML Support: Aligning linguistic quality with technical ML requirements. Multilingual Data Quality (EN–ES): Dataset review and validation. Semantic Consistency: Edge-case resolution and guideline development. QA Frameworks: Metrics, auditing, data integrity, and maintenance. MTPE & Linguistic QA: Post-editing and translation QA. Training & Enablement: Guideline onboarding, calibration, and coaching for consistent delivery. Cross-functional collaboration with AI Research and Engineering. PROFESSIONAL EXPERIENCE micro1 | Human Data Lead (HDL) Remote | 2025 – Present • • • • • Led large-scale human data operations, ensuring critical support for training pipelines. Optimized annotation/evaluation workflows, aligning output with model objectives. Delivered high-quality human data refinement impacting model performance. Guided contributors on guideline interpretation and edge-case handling. Improved dataset consistency and semantic accuracy across EN–ES inputs. micro1 | Human Data Manager (HDM) Remote | 2025 • • • • Designed and implemented workflows for training datasets. Conducted output reviews to improve training signal quality. Supported rapid iteration cycles through structured human evaluation. Collaborated with AI research and operations teams in fast-paced environments. Self-employed | Freelance Translator, Editor & MTPE Specialist 2001 – Present • • • Localized 10M+ words in technical, healthcare, fintech, and public-sector domains. Maintained a consistent 90%+ QA evaluation score across major global platforms. Optimized MTPE workflows and contributed linguistic expertise to early-stage ML and automated translation systems. Blend (One Hour Translation) | Translator & Reviewer 2009 – Present • • Ranked among the top-performing linguists based on quality and turnaround metrics. Localized complex healthcare, technical, and structured content; ensured terminology consistency and compliance. TOOLS / TECHNOLOGIES CAT/Localization: SDL Trados Studio, memoQ, Wordfast, Smartcat, Crowdin. Design/Docs: Adobe InDesign, Illustrator, Photoshop, Acrobat Pro. Collaboration/Ops: Slack, Google Workspace (Docs, Sheets, Drive). Productivity: Microsoft Office (advanced), web research. EDUCATION • • Degree in Technical, Scientific and Literary English-Spanish Translation | Lenguas Vivas, Argentina (GPA 8.8/10) Postgraduate University Teaching Degree (in progress) | Universidad de la Ciudad de Buenos Aires LANGUAGES • • • • Spanish | Native English | Near-native (C1–C2) Portuguese | Basic Italian | Basic PROFESSIONAL AFFILIATIONS • • Member | Argentine Association of Translators and Interpreters (AATI) Member | Fédération Internationale des Traducteurs (FIT)