LLM model trainer for a medical scoring system

About OpenTrain

OpenTrain aggregates data-labeling and AI-training jobs from many companies and platforms so you can find this kind of work in one place. Creating an OpenTrain account is free and applying takes only a few minutes.

We list contractor-friendly projects across domains, including medical annotation and LLM training tasks, so contributors can discover and apply quickly without hunting multiple sites.

Why AI training work matters in healthcare

AI training (also called data labeling, annotation, or human feedback work) is the human side of building clinical AI systems. People prepare and verify examples—like extracting clinical measurements—that models learn from.

This work is often remote, flexible, and accessible: many medical projects value attention to detail and domain familiarity. Contributors directly shape how models behave in sensitive medical settings.

• Fully remote: do this work from anywhere with an internet connection.
• Flexible: most projects let you choose your own hours and workload.
• Meaningful: your labels directly influence model safety and accuracy in healthcare.

The role

OpenTrain AI is hiring a contractor to collect specific clinical datapoints from a medical database to support training of an LLM-based scoring model.

You will use Med PALM 2 or BioBERT tools to extract urine output and serum creatinine datapoints from the MIMIC-4 database and prepare those examples so they can be used to train a model that predicts the occurrence of a clinical event.

• Project type: data collection for LLM training (label type: DATA_COLLECTION).
• Data type: MEDICAL_DICOM (medical database work tied to MIMIC-4).
• Labeling software: OTHER (use Med PALM 2 or BioBERT as specified).
• Employment: contractor; worldwide / remote.

What you'll do

Follow the project instructions to locate and extract the required clinical measurements from MIMIC-4 using the provided tooling (Med PALM 2 or BioBERT).

Format and document the extracted datapoints so they are ready for model training and quality review.

• Locate urine output and serum creatinine values in MIMIC-4 and extract them to the required schema.
• Use Med PALM 2 or BioBERT to assist extraction/annotation workflows as directed.
• Perform basic quality checks and document provenance for each extracted datapoint.
• Deliver organized data files and brief annotation notes for use in LLM training.

Requirements

This is an intermediate-level contractor role. You must be able to work directly with the MIMIC-4 database and with the specified LLM tooling (Med PALM 2 or BioBERT).

Prior experience training LLM models for medical projects is preferred but not mandatory; applicants without prior medical-LLM experience who can demonstrate competence with MIMIC-4 and the tooling are welcome to apply.

• Experience level: Intermediate.
• Required: ability to extract urine output and serum creatinine datapoints from MIMIC-4 using Med PALM 2 or BioBERT.
• Preferred: prior experience in medical LLM training projects (not required).
• Work arrangement: contractor, remote, worldwide.

Compensation and project details

This is a fixed-price contractor project: total payment is USD 1,000 for the deliverables specified in the assignment.

The project focuses on data collection (DATA_COLLECTION) from medical data (MEDICAL_DICOM) and uses non-standard/other labeling software as listed; you will be responsible for delivering the extracted datapoints in the requested format.

• Payment type: Fixed price — USD 1,000 for the project.
• Label type: DATA_COLLECTION. Data type: MEDICAL_DICOM.
• Labeling software: OTHER (Med PALM 2 or BioBERT specified).
• Employment type: CONTRACTOR; remote and open to applicants worldwide.

How to apply

Create a free OpenTrain account and apply through the listing. Applying takes only a few minutes.

When you apply, include a brief note describing your experience with MIMIC-4 and with Med PALM 2 or BioBERT, and indicate whether you have prior medical LLM training experience (preferred but not required).

• Submit: brief cover note, relevant experience summary, and any examples of similar work if available.
• Open to applicants worldwide; contractor engagement with a fixed-price deliverable.