Privacy Preserving Model Inference With Optional Data Retention Control

via “privacy-preserving model inference with optional data retention control”

AI-native code editor — Cursor Tab, Cmd+K editing, Chat with codebase, Composer multi-file.

Unique: Provides explicit privacy mode configuration that prevents code from being stored or used for training by model providers, addressing a key concern for enterprise users. Privacy setting is global and applies to all AI interactions in the editor.

vs others: More privacy-conscious than Copilot (which sends code to Microsoft/OpenAI by default) because it offers explicit opt-in privacy mode, but less transparent than local-only tools because the privacy mechanism is undocumented and still relies on cloud inference.

via “privacy-preserving memory storage with optional de-identification”

This package contains the code for training a memory-augmented GPT model on patient data. Please note that this is not the 'letta' company project with thehttps://github.com/letta-ai/letta; for use of their package, plsuse 'pymemgpt' instead.

Unique: Implements privacy controls as first-class memory operations rather than external post-processing; supports configurable de-identification policies that preserve clinical utility while protecting PII

vs others: More integrated than bolted-on privacy layers; privacy policies are enforced at memory storage level rather than just at query time

via “data-privacy-preservation-during-training”

via “privacy-preserving local inference”

via “privacy-compliant-predictive-modeling”

via “inference-time data access control and audit logging”

Unique: Applies attribute-based access control (ABAC) policies to inference requests, allowing rules like 'only users in department X can query model Y with data from region Z', rather than simple role-based access that doesn't account for data context

vs others: Provides inference-specific access control vs. generic API gateways (Kong, Apigee) which lack ML-specific policy semantics, and vs. model serving platforms (KServe, Seldon) which focus on performance rather than security audit trails