October 6, 2021
Today, I’m excited to share that I’ve joined Continual as Head of Marketing. Continual is radically simplifying the path to operational AI with the first continual AI platform built for the modern data stack. More in a bit on what that means, but the “so what?” is about opening the door for more organizations to embed AI across their business at scale. Think of Continual as a breakthrough in the proverbial wall to production: a next generation of ML platform codesigned with the modern data stack that finally democratizes operational AI.
Over my career, I’ve learned a lot helping customers make AI work for business, first with GPU optimized Power Systems at IBM, later with enterprise ML workflows at Cloudera, and most recently with ML-based fraud prevention solutions at HUMAN Security. Meanwhile, as the wall to making AI useful has moved further up the stack and deeper into the front lines of business, we’ve come a long way towards democratizing and making AI pervasive. Yet, despite the proliferation and maturation of DS/ML platforms, most organizations are still struggling to become truly AI-driven.
Software is eating the world, but data is its new super fuel. Turns out it’s hard to keep business applications and processes on a steady diet of data and analytical intelligence, and even harder to keep them continuously metabolizing ML predictions. If you’ve been anywhere near the ML application development lifecycle, you’ve tasted the practically infinite rainbow of failure modes. There are both the essential and incidental complexities, and then there are the walls — between infrastructure and tools, people and skills, processes and pipelines, and the gaps in all these — that get in the way.
Enter Continual, AI for the red hot modern data stack. I’m excited because Continual offers step-function gains over traditional ML/AI tooling in terms of cost, ease of use, and time to value, thanks much in part to being built to run where modern business data lives: on the cloud data warehouse. Continual erases the complexities of the ML development lifecycle by reducing the workflow to its essentials: specifying which data and features go in and what predictions come out. Users familiar with SQL, or dbt, can easily access this simplified workflow using a declarative analytics-as-code approach to building, deploying and maintaining continually improving predictions. These predictions are maintained in the data warehouse, accessible to downstream applications and modern data stack components like Census, Hightouch and more, giving data teams the ability to responsibly and programmatically feed continual predictions to real-world business applications.
Promises of leveraging machine learning for my own work as a B2B marketer have lived mostly in the realm of the imagination. Even while predicting which customers are about to churn has been technically feasible for years, doing that has mostly been up to whether or not we had the data engineering, data science and ML engineering skills available to work on that. Continual opens the door for every data professional to tackle AI use cases for the organization, cheaper and faster. With the rise of product-led, user-centric growth models in B2B, digital marketing and product experiences are becoming interconnected, opening up new ML use cases across a holistic digital customer experience. That triggers one’s imagination even more: for example, “Which content topics will product user A appreciate the most in their next email?”, or, “What early signals are predictive of customer lifetime value?”
I'm thrilled to join this team and can't wait to share some upcoming announcements with you soon. To see for yourself why I'm so excited about Continual, you can request a demo here.
Andrew Tsao discusses the main reasons he chose to join the Continual team.
Traditional AI platforms require users to manage complex data infrastructure and write bespoke ML pipelines for each case. We let users build predictive models using a declarative workflow that radically streamlines operational AI at scale.