Colin P.

Key Skills

Software

AWS SageMaker

Top Subject Matter

Clinical NLP

Medical AI

Healthcare Decision Support

Top Data Types

Text

Image

Document

Top Task Types

Fine-tuning

Prompt + Response Writing (SFT)

Freelancer Overview

Senior AI Full Stack Engineer (AI Data Labeling & Fine-tuning). Brings 9+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include AWS SageMaker, Internal, and Proprietary Tooling. Education includes Bachelor of Science, City University of Hong Kong (2018). AI-training focus includes data types such as Text and labeling workflows including Fine-tuning and Prompt + Response Writing (SFT).

Labeling Experience

Senior AI Full Stack Engineer (AI Data Labeling & Fine-tuning)

AWS SageMakerTextFine-tuning

Fine-tuned and deployed custom clinical NLP models on proprietary medical terminology data using AWS SageMaker in a healthcare setting. Evaluated and benchmarked foundation models on latency, clinical accuracy, and cost to optimize model selection. Developed retrieval-augmented generation (RAG) pipelines with Azure AI Search and pgvector, grounding LLM outputs in clinical knowledge bases and reducing hallucination rates. • Labeled and curated specialized medical datasets for fine-tuning language models in clinical decision support. • Implemented end-to-end prompt and output evaluation frameworks using Langfuse for continuous improvement and audit logging. • Applied RLHF and classification methodologies across real-world patient encounter data for model optimization. • Managed model versioning and A/B testing for improved inference quality and performance under HIPAA-compliant infrastructure.

2023 - 2025

Full Stack Engineer (LLM Data Labeling & Prompt Engineering)

TextPrompt + Response Writing (SFT)

Developed and maintained data pipelines for ingesting, normalizing, and storing physician-patient case histories and model-generated clinical text for use in LLM prompt training, output validation, and few-shot examples. Designed and implemented structured output validation and prompt evaluation frameworks with physicians to benchmark clinical completeness and accuracy of LLM outputs. Built core NLP and prompt engineering frameworks for structured reasoning tasks, supporting AI-generated differential diagnosis and medical documentation. • Engineered multi-step prompt workflows and function calling for accurate information extraction and structured response generation. • Curated and annotated datasets for evaluation, including clinical differential diagnosis cases, SOAP notes, and ICD-10 recommendations. • Implemented streaming LLM pipelines for real-time output data labeling and validation with clinical experts. • Collaborated with physician advisors to define clinical labeling guidelines and best practices for platform accuracy and safety.

2021 - 2023

Education

Work History