Stories

Accelerating AI-Driven Pricing Strategies with Outerbounds

Case Study image
400%

 Increase in Project Capacity: Went from managing 1 project to 5 projects simultaneously.

67%

Reduction in Overhead: Avoided hiring 2-3 engineers for infrastructure management.

5x

Faster Deployment: Increased velocity in building, training, and deploying ML models.

Name
Relu
Founded
2023
Location
Stealth mode, operating primarily online
Industry
E-commerce and Online Marketplaces
Focus on Machine Learning
AI-driven pricing strategies for e-commerce businesses.
ML Models
No items found.
No items found.

Relu, a UK-based startup, is carving a niche in the e-commerce sector by offering AI-driven pricing solutions. Led by Servando Torres García, an experienced data scientist serving as the Head of Data, Relu focuses on optimizing pricing strategies for e-commerce businesses. Their target customers are e-commerce founders with over three years of historical data, which enables Relu to implement data-driven strategies that enhance revenue and profitability.

Fragmented Tooling and Operational Overhead

Before adopting Outerbounds, Relu struggled with the complexities of managing their machine learning (ML) workflows. As the sole ML engineer at the company, Servando was responsible for everything from data preprocessing to model deployment. This was challenging, especially given the fragmented nature of the tools he had to work with. "I was investing a lot of time in the learning curve of new tools and assembling the logistics and operations,” Servando recalled.

Servando initially used a mix of tools such as GitHub Actions for CI/CD, Modal for cloud computing, and Gradio and Streamlit for front-end visualizations. This patchwork approach, however, led to inefficiencies and complications. “We were stuck in basically just putting together one project,” he noted, emphasizing how this setup slowed their progress and restricted their capacity to expand.

The fragmented tooling landscape not only consumed time but also made scaling operations a cumbersome process. Managing different tools and ensuring they worked together smoothly was a significant challenge, particularly for a small team. This scenario led Servando to search for a more streamlined solution that would allow him to focus on building and refining AI models rather than managing infrastructure.

Adopting Outerbounds for Scalable, Simplified Operations

Relu's search for a more cohesive solution culminated in the adoption of Outerbounds, a platform built on the open-source Metaflow framework. Servando first encountered Outerbounds through discussions in the machine learning community and was immediately impressed by its all-in-one capabilities. "When I saw the Outerbounds platform, my first impression was, ‘So they do the infra for me, and I do the ML?’ That was the trigger for us,” he explained.

Outerbounds provided exactly what Relu needed: an integrated platform that simplified the entire ML lifecycle. One of the key features that attracted Servando to Outerbounds was its seamless integration with Metaflow’s API, which streamlined the development and visualization of Directed Acyclic Graphs (DAGs). “Metaflow’s DAGs are the bridge between our business logic and the code implementation,” Servando said, noting how the platform made it easier for non-technical team members to understand the ML processes.

The user-friendly documentation and intuitive interface of Outerbounds also stood out. “The documentation walks you through the steps from testing locally to scaling and finally deploying to production. It’s a human touch that you don’t often find in software documentation,” Servando added.

Another critical feature was the simplicity of deploying ML models. With just a few lines of code, Relu could scale their models using cloud resources without worrying about the underlying infrastructure. This ease of use was a game-changer for Relu, allowing them to quickly adapt and scale their operations as needed.

Accelerated Development and Increased Project Capacity

Implementing Outerbounds has dramatically transformed Relu’s operations. One of the most significant improvements has been the increase in project capacity. Before adopting Outerbounds, Relu could only manage one project at a time. Now, they are handling five projects simultaneously—a 400% increase in capacity.

The reduction in engineering overhead has been another major benefit. Without Outerbounds, Relu would have needed to hire at least two or three additional engineers just to manage the infrastructure. "If we didn’t have Outerbounds, we’d need to hire at least three full-time employees to handle infrastructure, cloud management, and ML engineering," Servando estimated. Outerbounds has effectively eliminated this need, allowing Relu to allocate resources more efficiently.

The platform has also significantly increased the speed at which Relu can develop and deploy ML models. “Now, I can put together a prototype really fast, see some results, iterate on that, and still take care of stable projects. It’s the velocity that Outerbounds is providing for us,” Servando noted. This increased speed has enabled Relu to experiment with different models and approaches, leading to more effective AI-driven pricing strategies.

The Outerbounds UI has become an essential part of Relu’s operations, acting as their primary monitoring and reporting system. By leveraging the platform’s visualization tools, Relu can easily share key performance metrics with stakeholders, further streamlining their workflow and enhancing communication within the team.

Outerbounds as a Catalyst for Growth

Outerbounds has been instrumental in enabling Relu to focus on their core mission: delivering AI-driven pricing strategies that help e-commerce businesses optimize their profits. By reducing operational burdens and increasing development velocity, Outerbounds has allowed Relu to scale their operations without the need for additional engineering resources.

The platform’s all-in-one capabilities, combined with the powerful Metaflow framework, have improved Relu’s internal processes and made the company more attractive to potential hires. Servando believes that using cutting-edge technology like Outerbounds is key to attracting top talent. “If we tell our future engineers that this is the problem they’ll be solving, and that they’ll be using Metaflow and Outerbounds, it will be really appealing to them,” he said.

As Relu continues to grow, they are eager to explore all the features that Outerbounds has to offer, confident that the platform will continue to support their evolving needs. With Outerbounds handling the heavy lifting, Relu can concentrate on what they do best—innovating and delivering value to their customers through AI-driven solutions.