Real-time adaptability for your models

Orca helps machine learning teams update models instantly as business needs and data change, ensuring peak performance.

Talk to an ML Engineer

Models using Orca handle data evolution more effectively

accuracy loss for an image classification model experiencing severe data drift

20%

fewer errors for a sentiment classification scenario with shifting objectives

improvement in ROC AUC when detecting toxicity compared to a traditional classifier

Build AIs as dynamic as the real world

Instantly expand to new use cases or environments

Problem

Expanding to new use cases can be difficult and time-consuming due to the lack of a robust pre-existing data set.

Solution

Orca enables you to quickly build a new production-ready model using a small "golden" dataset. Then, as new data comes in, you can seamlessly optimize your model to improve performance without retraining.

Read blog post

Keep up with rapidly-evolving data

Problem

Large volumes of data often lead to data drift, which degrades AI/ML model performance, especially when data arrives faster than it can be processed

Solution

Orca uses a proprietary retrieval augmentation architecture that splits the model's knowledge from its reasoning capabilities. As a result, you can continually update your model's knowledge base to keep it performant even over long periods of time.

Read blog post

Customize to different preferences

Problem

AI/ML models often struggle to quickly adapt to diverse user requirements and changing definitions of "correct" outcomes due to sparse data availability and time-consuming retraining cycles.

Solution

Orca's unique architecture enables users to build performant models with sparse data and then continually update them as new data comes in. Its closed-loop system ensures that models remain accurate and responsive as definitions of "correct" outcomes change.

Read blog post

Adapt to changing business objectives

Problem

Businesses often change their strategic goals and priorities, which leads to their AI/ML models becoming outdated. Retraining or transfer learning can be inefficient and time-consuming.

Solution

Orca solves this by using a memory-augmented architecture that allows models to update in real-time with external data, enabling instant adaptation to new priorities.

Read blog post

How does Orca's augmentation methodology work?

Orca is an end-to-end system that integrates data management, models, and automated tuning for classification and other predictive systems. Orca teaches these models to leverage external data during fine-tuning, transforming them into "retrieval augmented" systems that adapt instantly to change. By separating reasoning from knowledge, models can be updated by modifying external data rather than retraining. Orca’s closed-loop nature also links model behavior to individual data points, enabling real-time data debugging and refinement, so you can build and update models with even just a few data points.

Read our documentation

from datasets import load_dataset
from orcalib.rac import LabeledMemoryset, RACModel

memoryset = LabeledMemoryset("datalicious") #cloud storage backed dynamic model memory dataset

source_dataset = load_dataset("datalicious/datalicious")
memoryset.insert(source_dataset) #works with all standard dataset formats

my_model = RACModel(num_classes=10)
with my_model.use(memoryset):
    my_model.finetune(source_dataset)


my_model.attach(memoryset)
my_model.predict(new_input)
memoryset.analyze() #tell me what could be better about my data

with my_model.use(new_memoryset):
    my_model.predict(new_input)