Appen

How we optimize processes and reduce infrastructure expenses through innovative solutions.

Industry

Data Labeling

Australia • China • Philippines • USA

Staff Augmentation

Deliverables

Data Engineering

AI/ML Solutions

Solution Architecture

Technologies

Kafka

Spark (Python)

FastAPI (Python)

Elasticsearch

AWS Redshift

The Customer

Appen is an Australian multinational company that provides high-quality data for machine learning and artificial intelligence applications. They offer various services such as data annotation, data collection, transcription, and translation to assist organizations in improving their AI systems and algorithms.

The Challenge

In the competitive landscape of data services for AI and ML, ensuring data integrity, client trust, and regulatory compliance is essential for maintaining a leadership position. Our challenge in this project involved refining the annotator registry process and implementing a robust anti-fraud measures to improve important business KPIs and reduce cloud infrastructure costs. 

Appen

35M

Records processed per hour

By integrating top-notch big data technology for data processing, we increase the number of events and parameters analyzed, reducing the time by 30%.

1M

In annual savings

Estimated right after the release of the new feature, including just cloud infrastructure and engineering savings.

5+

Years partnership

Successful nearshore software development partnership spanning over 5 years with multiple projects delivered.

01. Deepening our understanding of the persona, process, and problem.

We collaborated with the client’s business areas to gain a better understanding of the data sources and information flow. We identified potential fraud events throughout the process and defined classification mechanisms.

work-appen-team
work-appen-team

Understanding the business context and needs is crucial for proposing the most suitable architecture. In this instance, the solution needed to incorporate real-time fraud detection and be capable of processing millions of records per hour.

— Claudio Vasconcelos

Tech Lead @ Novatics

02. Co-creating alternative solutions and selecting the optimal strategy.

We collaborated with the client to develop an architecture capable of supporting the volume of data and meeting other technical and functional requirements. Additionally, we researched, trained, tested, and selected the most suitable AI models for event classification.

work-appen-team

03. Implementing the solution iteratively and incrementally

We delivered a machine learning-based qualification and prediction solution using Kafka and Spark, which optimizes the anti-fraud process by delivering real-time alerts.

work-appen-team-03

“Novatics helped the data science team prevent fraud by building necessary infrastructure and providing guidance on best practices for data engineering related to fraud detection.“

— Phoebe Liu

Data Scientist @ Appen

Want to understand more about the project, Talk to a Specialist .

Cyber Security

AI

Protexxa

How we helped Protexxa provide a cybersecurity solution for companies around the world.

Read Case Study

Model Catalog

ML

Verta

How our agile team made a complex solution feasible by creating a more user-friendly interface.

Read Case Study

Sports Platform

AI
ML

Coterie

How we helped Coterie launch a new digital product in just a few months.

Read Case Study

Do you need help to create your digital products?

Novatics is a consultancy firm specializing in software design and development. We support companies globally in delivering innovative projects on time and on budget.

Come Talk to Us!
Need help