Jocelyn

Key Skills

Software

Don't disclose

Clickworker

Top Subject Matter

LLM reasoning

prompt engineering

behavioral modeling

Top Data Types

Text

Document

Video

Top Task Types

RLHF

Question Answering

Classification

Data Collection

Transcription

Object Detection

Entity (NER) Classification

Segmentation

Evaluation/Rating

Freelancer Overview

Founder & Productivity Systems Developer (Self-Employed) — LLM workflow design and iterative evaluation (2022–Present). Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose. Education includes Bachelor of Science, Southern New Hampshire University (2021) and Bachelor of Business Administration, Maestro College (2022). AI-training focus includes data types such as Text and labeling workflows including RLHF, Question Answering, and Evaluation.

Labeling Experience

Logic-Based Taxonomy & Behavioral Signal Classification

TextEvaluation/Rating

Designed and implemented a proprietary "Signal Index" framework to categorize complex human interaction data. Responsibilities included developing a granular taxonomy for behavioral markers, performing multi-class intent classification, and establishing rigorous quality control protocols for high-variance communication datasets. This work focuses on distilling unstructured, high context human dialogue into machine interpretable logic models. Expertise includes iterative refinement of annotation guidelines, edge-case resolution, and ensuring high inter-annotator agreement through clear, rule-based logic mapping.

2026 - Present

Operations Manager (Indoor Plant Business) — structured communication quality control (2022–Present)

Don't discloseTextQuestion Answering

Managed high-signal customer communications by resolving complex queries with clear, concise explanations, mirroring processes used to tune AI responses. Designed procedures for consistency and accuracy in high-volume communication contexts, supporting evaluation-style rigor. Treated communication outcomes as a form of quality control analogous to training-time feedback loops. • Provided structured resolutions to complex customer questions. • Ensured consistency and accuracy via SOP-driven processes. • Used concise, helpful phrasing to improve outcome quality. • Applied a feedback and refinement mindset to recurring query patterns.

2022 - Present

Founder & Productivity Systems Developer (Self-Employed) — LLM workflow design and iterative evaluation (2022–Present)

Don't discloseTextRLHF

Founded and operated logic-based productivity and LLM workflows focused on improving model reliability and reducing hallucinations through iterative testing. Used behavioral analysis of user needs to structure prompt logic and drive more consistent, low-noise outputs for end users. Applied human-feedback-like principles by evaluating algorithmic outputs and refining them for clarity and integrity in real-world interactions. • Performed iterative evaluation of LLM outputs to minimize hallucination risk. • Modeled user behavior patterns to translate needs into structured logical sequences. • Refined inputs to prioritize relevance over verbosity. • Built tools intended to support response tuning through analogous RLHF practices.

2022 - Present

Data Verification Specialist

TextData Collection

Performed data validation to ensure the accuracy and integrity of web-based information. Executed content auditing to identify and rectify discrepancies in website data. Conducted relevance evaluation and fact-checking to support search engine algorithm training. Maintained high standards of data quality assurance by verifying information against strict project guidelines.

2017 - 2020

Research Assistant (Southern New Hampshire University) — evidence synthesis and data verification (2021)

Don't discloseText

Conducted quantitative analysis and evidence-based synthesis of complex environmental datasets and literature, emphasizing meticulous verification and reporting. Transformed scientific information into coherent, fact-checked outputs that demonstrate high-quality information processing. The work functionally resembles dataset validation and model-style evaluation where outputs must be consistent, accurate, and grounded in sources. • Performed careful analysis of complex environmental datasets. • Synthesized diverse literature into evidence-based reports. • Completed fact-checking and adherence to strict reporting standards. • Produced coherent summaries from heterogeneous scientific inputs.

2021

Education

Work History