Welcome to Snowflake’s Startup Spotlight, where we learn about the people and companies building businesses on Snowflake. In this Q&A with Fernando Carrasco, Co-Founder of Inventa and leader of Operations, Tech, and Data Strategy, we’ll learn why he’s inspired by small and medium businesses and how ML and large language models (LLMs) are helping Inventa reinvent the wholesale ecosystem for small retailers across Brazil.
What inspires you as a founder?
I have always been passionate about improving the productivity of small and medium businesses, which are at the core of the economy of most countries. As my father is a small business owner, I have been familiar with their challenges from an early age and always thought technology was a key lever to level the playing field.
My professional journey has centered around solving business problems by leveraging data and AI. This experience allowed me to develop some valuable entrepreneurial skills that I’ve been able to use in starting Inventa: how to discover client needs, develop a value-driven vision for a digital product, hire and empower an exceptional team, and leverage data and AI to serve our clients.
How does that inspiration translate into what Inventa does?
Small business owners across Latin America face many challenges today. Inventa is the online wholesale marketplace for small retailers and suppliers across Brazil. Our platform helps both sides of the wholesale ecosystem succeed. For retailers, this means recommending suppliers and products on our marketplace, extending credit so they can purchase what they need, and reducing shipping costs. Suppliers have a place to manage their entire wholesale business, drive more sales from both new and existing customers, and help them better understand their business with analytics.
What’s the coolest thing your company is doing with data?
We incorporated ML into our product from the very beginning. ML plays a role in everything from predicting delivery times (a much harder problem in Brazil than in the U.S.) and ranking our marketplace to delivering our search experience. Looking ahead, the widespread use of WhatsApp in our market will allow us to leverage LLMs to create familiar but automated purchase experiences for our retailers.
How does Snowflake fit into the Inventa stack? What role does it play?
Data engineering is always a bottleneck for data teams. This is especially true in Brazil where the data engineer role is still gaining traction. Snowflake allowed us to scale our data platform and data team without this blocker. Given our pace of growth, this has been crucial for data to keep up with the business.
The managed architecture allowed us to build user-facing data products quickly and reliably. This was the same reason we went with Snowflake to solve the original problem of needing a robust, low-maintenance data infrastructure for internal use cases. We could get up and running quickly and prevent data platform management from blocking development.
How does Snowflake enable you to push the envelope in your line of business?
We are not only running all of our ingestion, transformation, and analytics on Snowflake, but much of our ML infrastructure as well. Snowflake is a key piece of our ML feature platform along with Feast and Redis. We also use it for batch inference of our ML models. The capabilities of Snowpark make this integration seamless. We’re also excited to explore how Unistore can play a role in our ML stack in the future.
What’s the most valuable piece of advice you got about how to run a startup?
Running a startup is about hiring and coaching a great team and enabling them to work together while solving a relevant need for a clear target customer. Marcos [Salama, CEO and Co-Founder of Inventa] and I got that advice from Daniel Scandian, CEO and Founder of MadeiraMadeira, a very successful, consumer-focused home decor marketplace.
Let’s explore the flip side of that question: What’s a valuable lesson you learned the hard way?
I’ve learned that tech and data migrations are difficult and can become messy if they are not being very carefully planned. They also require clear governance for coordinating the different tech teams involved (Product Management, Engineering, Data, and Design).
When you envision the future of your industry, where do you see Inventa?
Inventa is in a position to eventually become the ecosystem of B2B wholesale in Brazil: the go-to-platform for any supplier (both brands and distributors) that want to optimally grow and manage their sales to retailers. We will help them integrate all their channels, solve their working capital needs, and automate their sales and post-sales back-end processes.