Jes L.

Key Skills

Software

Labelbox

Scale AI

Prodigy

AWS SageMaker

Top Subject Matter

Digital Products & SaaS

Media Intelligence & Content Analysis

Human Behavior & Decision-Making

Top Data Types

Text

Image

Document

Top Task Types

Data Collection

Segmentation

Classification

Text Generation

Question Answering

Text Summarization

Evaluation/Rating

RLHF

Object Detection

Entity (NER) Classification

Freelancer Overview

My experience with AI training data is grounded in human-centered design, qualitative research, and operational strategy rather than traditional “labeling at scale,” which I see as a strength. I’ve spent the past several years translating complex human behaviors, contexts, and decision-making patterns into structured insights that inform intelligent systems. In roles such as Strategy Director at Stellar Elements and Director of Design Research at Cision, I led large-scale research initiatives where we defined taxonomies, categorized unstructured data, and built frameworks that effectively “labeled” human experiences—turning ambiguity into usable training inputs for product, content, and AI-driven solutions. This includes developing classification schemas, annotating qualitative data, and ensuring consistency and rigor across datasets used to inform automation, personalization, and AI-enabled products. What sets me apart is my ability to bridge the gap between raw human input and meaningful data structures. I bring deep expertise in ethnographic research, bias identification, and sensemaking—critical for creating high-quality training data that reflects real-world nuance rather than oversimplified categories. I’ve also worked closely with cross-functional teams to operationalize these frameworks, ensuring they scale and remain aligned with business goals. My background allows me to approach data labeling not as a mechanical task, but as a design challenge—one that requires clarity, ethical consideration, and a strong understanding of how data shapes model behavior.

Labeling Experience

LLM Response Evaluation & Prompt-Response Annotation

TextRLHF

Evaluated and annotated AI-generated outputs for quality, relevance, and alignment with user intent. Compared multiple responses and ranked them based on clarity, usefulness, and accuracy, similar to reinforcement learning from human feedback (RLHF) workflows. Provided structured feedback to improve model outputs and ensure alignment with human expectations. Applied human-centered evaluation criteria to capture nuance, edge cases, and contextual appropriateness in generated content.

2024 - Present

Behavioral Annotation & Journey Classification for Digital Products

TextClassification

Annotated user behaviors and interactions across customer journeys by labeling key stages, actions, and outcomes. Classified behavioral patterns such as intent, drop-off points, and conversion triggers. Used evaluation frameworks to assess experience quality and consistency across touchpoints. Ensured labeling accuracy by defining clear criteria and resolving ambiguous interaction patterns. Outputs informed service design improvements and predictive modeling efforts.

2005 - Present

STOA — AI Governance System (Research & Development)

Don't discloseTextData Collection

Led research and development of an AI governance system focused on building transparent, user-controlled context through a protocol and prototype. The work involved exploring requirements and evidence for making social media content understanding more interpretable to users. Deliverables included a research prototype intended to guide how context and labeling-like governance information could be represented. • Built the research plan, prototype, and protocol for transparent context. • Focused on user-controlled context to explain how and why beliefs are shaped. • Produced a preprint documenting the approach. • Designed for transparency and rights-aligned governance of digital interactions.

2024 - 2025

Taxonomy & Ontology Design for Financial Services Data Structuring

TextClassification

Designed and implemented structured taxonomies to classify financial products, user needs, and decision criteria. Labeled and organized complex, unstructured inputs from stakeholder interviews and documentation into consistent entity categories and hierarchical schemas. Translated ambiguous concepts into clearly defined labels that could be applied consistently across datasets. This work supported structured data modeling and improved the usability of data for downstream systems and decision-making.

2023 - 2023

Qualitative Data Annotation for User Intent & Sentiment Classification

TextClassification

Led large-scale annotation of unstructured qualitative data, including user interviews, customer feedback, and support transcripts. Developed and applied a structured coding framework to label user intent, sentiment, pain points, and decision drivers. Created detailed annotation guidelines to ensure consistency across contributors and reduce ambiguity in edge cases. Implemented iterative taxonomy refinement based on observed inconsistencies and emerging patterns. Conducted quality checks and alignment sessions to improve inter-annotator agreement. The resulting labeled dataset informed product strategy, personalization features, and AI-driven content recommendations.

2021 - 2022

Education

Work History