Neptune Technologies Inc.

Effective Date: January 1, 2025

Neptune Technologies Inc. ("Neptune," "we," "our") is committed to training AI systems responsibly, with strict attention to privacy, transparency, and control. This Notice explains how our AI models are trained and how data used in training is handled and safeguarded.

1. Overview of Neptune’s Models

1.1 Enterprise Models
AI systems trained solely on proprietary customer data in a secure, isolated deployment (on-premises or private cloud). These models are custom-tailored to client needs and never used outside their designated application.

1.2 Consumer-Facing Models
General-use models trained only on publicly available, curated, domain-specific datasets (e.g., financial news, stock data). No personal user data is used for training unless we obtain explicit, informed consent.

2. How Models Are Trained

2.1 Pre-Training
Model training begins with pre-approved datasets. Enterprise models are trained solely on customer-provided data, isolated from Neptune systems. Consumer models use non-personal, de-identified sources.

2.2 Fine-Tuning
We refine models through human feedback (RLHF or supervised reviews). For enterprise models, fine-tuning occurs only if authorized and only using data the customer has explicitly permitted for such use. No enterprise models are ever shared.

2.3 Free Tier and Model Improvement
Data submitted through free-tier services may be used to improve our systems, but never without full transparency. Users can opt out through account settings or by contacting privacy@neptunetechnologies.com.

3. Data Sources and Permissions

Enterprise Customers: Own and control their data. Training is performed in a secure, logically segmented environment. We require written agreements confirming permission for training use.

Consumer Data: Only public, licensed, or open-access datasets are used. No personal identifiers are included.

4. Privacy, Security, and Safeguards

• All data undergoes de-identification before training unless specifically required otherwise and approved by legal contract.

• We enforce strict access controls, logging, and audit trails.

• All personnel involved in training are subject to confidentiality and security training.

• Neptune complies with applicable regulations including GDPR, CCPA, and LGPD.

5. Safety Commitment

We incorporate responsible AI practices to prevent harmful, biased, or unethical behavior. Model behavior is continuously tested against safety benchmarks. We prohibit the use of training data that may introduce discriminatory or unlawful behavior.

6. Your Rights and Contact

If you wish to request access, deletion, or information, contact us at:

privacy@neptunetechnologies.com

Neptune Technologies Inc.

Last Updated: January 1, 2025