Unlocking Data Bottleneck: The Journey of Gretel

Unlocking Data Bottleneck: The Journey of Gretel

Table of Contents

1. Introduction
2. The Genesis of Gretel
3. Identifying the Data Bottleneck
4. Building Gretel: Early Conversations and Foundations
5. Advantages of Synthetic Data
6. Expanding the User Base: From Data Engineers to Individuals
7. Gretel in Healthcare
8. Onboarding Efforts and Working with Research Institutions
9. Applications in Web 3.0 and Gaming
10. Future Roadmap: Becoming the Single Platform for Synthetic Data
11. Investing in Usability and Integration
12. Company Growth and Hiring Plans

Introduction

Welcome to this episode of "Gray Matter," the Podcast from Greylock where we share stories from company builders and business leaders. In this episode, I had the pleasure of speaking with Ali Gulshan, CEO and co-founder of Gretel AI. Gretel AI is often referred to as the "GitHub of data." But what exactly does that mean? In this interview, Ali discusses the journey of building Gretel, the challenges they set out to solve, and the unique advantages of synthetic data. We also dive into the application of Gretel in various industries, from healthcare to web 3.0. So let's jump right into it and learn more about the world of synthetic data and the future of data engineering.

The Genesis of Gretel

Ali starts by explaining the background of the three co-founders of Gretel, who came from the intelligence community and had firsthand experience of the challenges in working with data. They noticed the emergence of walled gardens around data, which provided a massive advantage to larger companies. This led to the idea of democratizing access to data and building comprehensive tools that could be accessed and used by everyone. Inspired by GitHub's impact on code collaboration and democratization, they set out to remove the data bottleneck and make data more accessible.

Identifying the Data Bottleneck

Ali explains that just as the compute bottleneck was a challenge for engineers in the past, there is now a data bottleneck that hinders the progress of developing applications and services. The higher entropy of data, along with ethical, privacy, and compliance concerns, makes it difficult to access and use data effectively. Gretel's goal is to unlock and remove this data bottleneck by building an ecosystem that provides high-quality synthetic data, rapid development, and collaboration tools. They aim to empower organizations of all sizes to build with data and leverage ML/ai technologies without the need for extensive expertise.

Building Gretel: Early Conversations and Foundations

Ali shares that the founding team of Gretel has a background in the intelligence community and building companies that were eventually acquired. Their experience spanned from starting with a lack of data and resources to working in organizations with abundant data. Combining their expertise and the growing trend of restricted data access, they identified the need for a platform like Gretel. Early conversations with investors, including Greylock, revolved around the main questions users and customers needed answers to: data quality, use cases, and privacy. These conversations helped Shape the product roadmap and validate Gretel's potential as a company.

Advantages of Synthetic Data

Ali emphasizes the advantages of using synthetic data compared to raw data. Synthetic data can often yield better results as it allows for better labeling, transforms, and privacy. It solves the challenges of biased and lower-quality raw data sets. Ali cites a study conducted by Gretel's CPO, Alex, which demonstrated that synthetic data can produce better results in many cases. The comprehensive tooling provided by Gretel allows users to easily generate high-quality synthetic data, improving the accuracy and reliability of ML/AI models.

Expanding the User Base: From Data Engineers to Individuals

Initially, Gretel focused on personas like data engineers, developers, ML/AI engineers, and data scientists. However, they later discovered that their usability and ease of use attracted a much broader user base. They saw individuals, startups, SMBs, and smaller companies using Gretel's low-code/no-code approach to data engineering. Users could easily synthesize, label, or transform their data without requiring extensive expertise. This ease of use and democratization of data engineering was a key factor in Gretel's mission.

Gretel in Healthcare

Ali highlights the significance of healthcare and health sciences as an area of focus for Gretel. The COVID-19 pandemic accelerated the importance of data-driven decision-making in healthcare. Partnerships with health sciences organizations, like Illumina, allowed Gretel to work on synthetic data sets of genotypes and phenotypes, enabling hospitals and researchers to benefit from high-quality research data. Gretel's tools also helped improve the detection of female heart disease and generate variations of skin anomalies. The goal is to leverage synthetic data to remove biases, improve the accuracy of predictions, and facilitate better healthcare outcomes.

Onboarding Efforts and Working with Research Institutions

Ali explains that Gretel built its product to be a self-serve platform for engineers, developers, and scientists. The goal is to automate complexity and make data engineering accessible to everyone. While strategic partnerships with research institutions and healthcare organizations are critical for Gretel, the majority of users can sign up on the website, drop their data in the console, and start synthesizing, classifying, or transforming data immediately. Gretel aims to facilitate rapid onboarding without requiring extensive training or complex integrations.

Applications in Web 3.0 and Gaming

Ali discusses the emerging applications of Gretel in Web 3.0 companies and gaming. They've seen increased interest from companies trying to bring financial services to Web 3.0 and needing better forecasting and testing capabilities. Additionally, some gaming companies rely on synthetic data to replicate production chains in test networks. By training on pseudo-randomly generated data, they can make better predictions for new models and in-Game variations. These use cases demonstrate the versatility of synthetic data and its potential across various industries.

Future Roadmap: Becoming the Single Platform for Synthetic Data

Ali envisions Gretel becoming a single platform for all types of synthetic data. The goal is to combine different data types, such as tabular, image, and visual data, to make accurate predictions. Gretel aims to be a comprehensive platform that enables users to build, train, and make predictions on synthetic data. They also emphasize the importance of reporting and showcasing the efficacy of synthetic data compared to raw data. By providing deep visibility and correlated insights, Gretel aims to be a trusted platform for data engineering and predictive analytics.

Investing in Usability and Integration

Ali discusses Gretel's focus on usability, automation, and integration. They want to make Gretel's toolkit easy to use and integrate with existing systems and tools. The goal is to remove complexity and make data engineering accessible without requiring extensive resources or expertise. Gretel is investing in building connectors to popular storage systems and integration with Apache Airflow for streamlined synthetic data pipelines. The aim is to become a seamless part of the data engineering ecosystem and automate end-to-end workflows.

Company Growth and Hiring Plans

Ali shares that Gretel is currently a fully distributed and remote-first company with a team of 40 people across the US and Canada. They plan to double in size over the next year and continue growing rapidly. Gretel is hiring across multiple departments, including engineering, applied research, marketing, sales, customer success, and solutions. The focus is on building a team that aligns with their vision of usability, automation, and making data accessible to everyone. Gretel values creating a people-friendly organization and is committed to fostering an environment of innovation and collaboration.

Conclusion

In this interview with Ali Gulshan, CEO and co-founder of Gretel AI, we explored the evolution of Gretel, the advantages of synthetic data, and its applications in various industries. Gretel's mission is to remove the data bottleneck and make data more accessible and usable for developers, engineers, and scientists. With a focus on usability, integration, and comprehensive reporting, Gretel aims to become the single platform for all types of synthetic data. As they continue to grow and expand their user base, Gretel is making data engineering more accessible and accelerating the adoption of synthetic data to drive innovation and better decision-making.