Financial Chatbot Labeler, PFM, Function-Calling & SQL

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We connect people with real projects that help shape how modern AI systems behave and offer accessible, flexible remote work.

About AI Training Work

AI training (also called data labeling or annotation) is the human work that teaches models how to understand and respond. Tasks range from classifying intents and labeling entities to evaluating assistant responses and validating structured calls to backend APIs.

This project focuses on conversational and function-calling data for a personal financial assistant — a high-impact area where your annotations directly improve user experience and model safety.

The Role

You will annotate a dataset of financial conversations intended to fine-tune and evaluate an AI financial advisor. Each sample becomes a structured JSON record with fields for the user query, labeled intent, function name and arguments (when applicable), assistant reply, and SQL-safety judgments.

• Employment type: Contractor.
• Platform: Annotations are completed in AWS SageMaker.
• Data type: Text — conversational banking queries and assistant replies.
• Label types: Classification, fine-tuning guidance, function-calling tags, QA, and text-generation labels.

What You'll Do

Annotators will review conversation samples and produce structured JSON outputs that capture intent, transaction/merchant categories, function-call usage, and assistant response quality. You will also mark whether SQL queries are safe, parameterized, and read-only.

• Classify user intent (examples: spending insight, budgeting, card support).
• Label transaction and merchant categories when present in the conversation.
• Tag function calls and map to provided function names (get_user_summary, get_transactions, run_custom_sql).
• Evaluate assistant replies for tone, completeness, and professionalism.
• Validate SQL: mark queries as safe/unsafe and note parameterization and read-only status.
• Produce complete structured JSON records with query, labels, function name/arguments, and assistant reply.

Requirements

This is an intermediate-level annotation project that requires domain familiarity and practical labeling experience. All required qualifications below come from the project brief and are mandatory to perform the work accurately.

• Familiarity with personal finance terminology and common banking concepts.
• Experience labeling chatbot or conversational data (intent tagging, response evaluation).
• Understanding of function-calling or API-style structured data and how calls map to backend actions.
• Basic SQL literacy sufficient to identify whether a query is safe, parameterized, and read-only.

Who Should Apply

Apply if you have hands-on annotation experience with conversational agents and a working knowledge of personal finance. This project suits contributors who can balance accuracy and consistency while following labeling guidelines in AWS SageMaker.

• Intermediate annotators with prior chatbot labeling experience.
• People comfortable with structured outputs (JSON) and careful quality checks.
• Annotators available to work remotely from any country (worldwide).

How It Works & Compensation

This project is a fixed-price contractor engagement paid at $2,000 for the completed assignment. Work is performed remotely in AWS SageMaker following detailed labeling guidelines provided by OpenTrain.

All annotations must be delivered as structured JSON records per sample. No minimum hours or fixed schedule are required; you will follow the project timeline and submission instructions supplied after onboarding.

• Payment: Fixed-price $2,000 (USD) for the project.
• Platform: AWS SageMaker for annotation tasks and submissions.
• Location: Remote — contributors worldwide may apply.
• Onboarding: You will receive labeling guidelines, examples, and validation checks before starting.