Kevin Portuondo

Key Skills

Software

Scale AI

Top Subject Matter

LLM Training

Audio AI

Multimodal AI

Top Data Types

Audio

Video

Text

Document

Top Task Types

RLHF

Prompt + Response Writing (SFT)

Freelancer Overview

Director of AI Operations & Multimodal Data Programs. Brings 3+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Internal, Proprietary Tooling, and Scale AI. Education includes Associate of Arts Degree, Florida State University and Honors Diploma, Saint Thomas Aquinas. AI-training focus includes data types such as Audio, Video, and Text and labeling workflows including Evaluation, Rating, and RLHF.

Labeling Experience

Director of AI Operations & Multimodal Data Programs

TextRLHF

Managed large-scale RLHF and SFT-based LLM training data projects, including prompt engineering and golden dataset creation. Authored operational standards converting engineering specifications into labeling SOPs and QA rubrics. Oversaw the ongoing operation, scaling, and delivery of hundreds of thousands of labeled data points for foundation models. • Coordinated LLM data operations spanning 10+ live workflows • Promoted and developed 20+ project leads to expand leadership pipeline • Defined project guidelines, review systems, and benchmarks for LLMs • Ensured delivery of high-quality labeled text data to AI teams

Managed large-scale RLHF and SFT-based LLM training data projects, including prompt engineering and golden dataset creation. Authored operational standards converting engineering specifications into labeling SOPs and QA rubrics. Oversaw the ongoing operation, scaling, and delivery of hundreds of thousands of labeled data points for foundation models. • Coordinated LLM data operations spanning 10+ live workflows • Promoted and developed 20+ project leads to expand leadership pipeline • Defined project guidelines, review systems, and benchmarks for LLMs • Ensured delivery of high-quality labeled text data to AI teams

2022 - 2024

Director of AI Operations & Multimodal Data Programs

Video

Oversaw the delivery and evaluation of video datasets for AI training, QA, and benchmarking. Managed and scaled teams for pipeline quality control and resolved project-level escalations and throughput concerns. Developed guidelines to improve data accuracy and consistency for video-based foundation models. • Led 100+ QA team members across concurrent video workflows • Implemented structured review and performance systems • Delivered consistent quality for high-volume video data labeling • Coordinated ramp of simultaneous multi-project video pipelines

Oversaw the delivery and evaluation of video datasets for AI training, QA, and benchmarking. Managed and scaled teams for pipeline quality control and resolved project-level escalations and throughput concerns. Developed guidelines to improve data accuracy and consistency for video-based foundation models. • Led 100+ QA team members across concurrent video workflows • Implemented structured review and performance systems • Delivered consistent quality for high-volume video data labeling • Coordinated ramp of simultaneous multi-project video pipelines

2022 - 2024

Director of AI Operations & Multimodal Data Programs

Audio

Led the creation and scaling of multimodal AI data verticals, delivering over 250,000 data points and 25,000+ audio hours to support foundation model training and evaluation. Designed and implemented benchmarks, QA frameworks, and review layers across 25+ AI projects to ensure quality and taxonomy consistency. Established structured performance tracking systems and guidelines for large-scale annotation and validation workflows. • Scaled contributor base from 0 to 500+ for multimodal data projects • Authored SOPs and standards for data project execution • Drove real-time framework iteration and operational ramp • Served as operational owner coordinating between US and China-based engineering

Led the creation and scaling of multimodal AI data verticals, delivering over 250,000 data points and 25,000+ audio hours to support foundation model training and evaluation. Designed and implemented benchmarks, QA frameworks, and review layers across 25+ AI projects to ensure quality and taxonomy consistency. Established structured performance tracking systems and guidelines for large-scale annotation and validation workflows. • Scaled contributor base from 0 to 500+ for multimodal data projects • Authored SOPs and standards for data project execution • Drove real-time framework iteration and operational ramp • Served as operational owner coordinating between US and China-based engineering

2022 - 2024

Senior QA Lead & High-Priority Intervention

Scale AIText

Served as Senior QA Lead on multimodal data labeling projects, reviewing and evaluating outputs against enterprise benchmarks. Developed and delivered testing protocols, quality standards, and rubrics for model performance and consistency. Audited, corrected, and provided direct feedback to contributors on 1,000+ tasks weekly. • Supported large annotation teams (up to 5,000 contributors) • Standardized project-wide evaluation criteria and rubric rewrites • Ensured readiness of production data for high-profile AI clients • Drove process improvements to stabilize annotation workflow

Served as Senior QA Lead on multimodal data labeling projects, reviewing and evaluating outputs against enterprise benchmarks. Developed and delivered testing protocols, quality standards, and rubrics for model performance and consistency. Audited, corrected, and provided direct feedback to contributors on 1,000+ tasks weekly. • Supported large annotation teams (up to 5,000 contributors) • Standardized project-wide evaluation criteria and rubric rewrites • Ensured readiness of production data for high-profile AI clients • Drove process improvements to stabilize annotation workflow

2021 - 2022

AI Prompt Engineer

TextPrompt Response Writing SFT

Engineered and rewrote prompts for LLM-based conversational AI systems, directly shaping the assistant's personality and engagement style. Evaluated and iterated on prompt design to enhance response generation quality, tone calibration, and creativity. Delivered hundreds of high-quality prompt/response tasks under strict project timelines. • Focused on user experience optimization in large language models • Independently contracted, maintaining rigorous daily output • Drove improvements for LLM interaction paradigms and creative intent • Delivered contract deliverables efficiently for high-volume prompt generation

Engineered and rewrote prompts for LLM-based conversational AI systems, directly shaping the assistant's personality and engagement style. Evaluated and iterated on prompt design to enhance response generation quality, tone calibration, and creativity. Delivered hundreds of high-quality prompt/response tasks under strict project timelines. • Focused on user experience optimization in large language models • Independently contracted, maintaining rigorous daily output • Drove improvements for LLM interaction paradigms and creative intent • Delivered contract deliverables efficiently for high-volume prompt generation

2019 - 2021

Education

Work History