More model deployments per year
Less time to onboard a new data scientist
Less time to retrain models
About Tala
Tala is a financial services platform and micro-lender in the Philippines, Mexico, and Kenya, serving historically underserved populations. It has been a for-profit company dedicated to social impact since its founding, and its model depends on scale. Will High, Head of Data Science and Machine Learning Engineering, leads the team that runs Tala’s real-time credit approval models. Those models make decisions in seconds.
“Shivani Siroya, our founder, made an intentional choice early on to pursue a for-profit model, believing it would allow us to build a generational company that could ultimately serve more people than we ever could as a nonprofit,” Will explains. “To achieve the scale we’re aiming for, we leverage nontraditional data and innovative DS-based underwriting, performing real-time credit assessments on Android devices,” which dominate mobile usage in Tala’s markets.
Growing in emerging markets takes infrastructure that automates machine learning pipelines. Tala found it in AI orchestration, the Metaflow-based capability of the Anaconda Platform that runs machine learning workflows in production.
Retraining by Hand Took Months
Will’s team had tried Kubeflow and Airflow and found both burdensome.
“Kubeflow was hard to manage. It required too much configuration and too many low-level commands,” says Will.
Models were retrained by hand in Jupyter notebooks, which took months. Updating a model often broke the code, and every break delayed the next release. In a business that calculates creditworthiness in real time, models have to stay current.
Onboarding was just as slow. “We used to have engineers set up EC2 instances and run Jupyter servers manually, which would take months to onboard a new data scientist,” Will notes.
The setup also couldn’t keep up with growing data volumes and a growing customer base. Tala needed something better.
“I’ve been using Metaflow since before it was open-sourced, but we needed more—something that would give us the scalability and compliance we required,” says Will.
After testing the open-source version, Will’s team realized they needed a more managed solution. Enter AI orchestration in the Anaconda Platform, which provided an enhanced version of Metaflow tailored for enterprises. The SLA-backed services, intuitive user interface, and compliance features were pivotal in Tala’s decision to choose the Anaconda Platform.
“The Anaconda Platform offered a UI that we loved and a strong SLA, which was critical for our operations. The automation it enabled was a game-changer,” Will recalls.
Why Tala Chose Managed Metaflow
Tala already knew Metaflow. Will had used it at Netflix since before it was open-sourced, and his team tested the open-source version against MLflow and Kubeflow.
“I’ve been using Metaflow since before it was open-sourced, but we needed more, something that would give us the scalability and compliance we required,” says Will.
Open-source Metaflow alone fell short, and Tala needed a managed service. That is the platform Outerbounds built on Metaflow, now offered as AI orchestration in the Anaconda Platform. It adds a service-level agreement (SLA), an interface for monitoring workflows, and built-in compliance features.
All of this “offered a strong SLA, which was critical for our operations. The automation it enabled was a game-changer,” Will recalls.
What AI Orchestration Is, and How It Fits With Anaconda
AI orchestration runs AI and machine learning workflows in production. Data scientists write workflows in Python. AI orchestration runs them, automates model retraining, and records every run and artifact, all on the cloud infrastructure a team already uses.
It is built on Metaflow, the open-source framework created at Netflix, where it supports 3,000+ AI and ML projects. Thousands of companies use Metaflow. Outerbounds, the company behind it, built the enterprise platform on top. Anaconda acquired Outerbounds in April 2026, and AI orchestration is now the Operate capability of the Anaconda Platform. Anaconda continues to contribute to Metaflow as an open-source project.
Teams can start with AI orchestration on its own, then connect it to trusted packages and curated models from Anaconda.
Five Times the Deployments, Onboarding in Days
With AI orchestration in place, model retraining runs automatically on fresh data, with no manual notebook work.
“We used to spend months fixing broken notebooks; now, with the Anaconda Platform, it’s all automated. The agility we gained is incredible,” says Will. That automation lets the team handle the data it processes daily as it enters new markets.
New data scientists now get set up in days instead of months. AI orchestration works with Tala’s existing cloud environment on Amazon Web Services (AWS), and its compliance features help Tala meet financial services regulations.
“Our onboarding process for new data scientists went from months to just days. That’s a huge win for us,” Will emphasizes.
Deployments climbed from a few models a year to more than five times that, which means more accurate credit decisions and better service for customers. The interface also lets the team monitor workflows in real time and debug them quickly.
“We now have five times the model deployments we had before, and the impact on our business is clear: we’re reaching more people, faster,” adds Will.
Reproducibility turned out to matter just as much. Tala’s data scientists can revisit any model run from the past two years, which supports compliance and audits of credit decisions.
“Reproducibility is the sleeper feature. You don’t realize how important it is until you need it. The Anaconda Platform has made it easy to access every run and artifact from the past two years,” Will explains.
With AI orchestration, Tala:
- Cut model retraining time by 95%: Automated retraining replaced manual notebook work that took months.
- Deployed five times more models: More deployments support real-time credit decisions.
- Reduced onboarding time by 90%: New data scientists go from months of setup to days.
- Gained reproducibility for compliance: Any run and artifact from the past two years is available for audits.
It is built on Metaflow, the open-source framework created at Netflix, where it supports 3,000+ AI and ML projects. Thousands of companies use Metaflow. Outerbounds, the company behind it, built the enterprise platform on top. Anaconda acquired Outerbounds in April 2026, and AI orchestration is now the Operate capability of the Anaconda Platform. Anaconda continues to contribute to Metaflow as an open-source project.
Teams can start with AI orchestration on its own, then connect it to trusted packages and curated models from Anaconda.
What’s Next
Tala plans to bring AI orchestration to new countries and explore AI-driven user experiences with graphics processing unit (GPU)-powered inference. That foundation positions Tala to keep growing while it pursues its mission of financial inclusion.
“We are building something truly scalable, and Anaconda has been critical in helping us achieve that,” concludes Will.
Want to scale AI from development to production on your own terms? Contact our team for a personalized demo.
Want to scale AI from development to production on your own terms? Contact our team for a personalized demo.
AI Orchestration
Discover how the Anaconda Platform can transform your data science workflows, while maintaining enterprise-grade security and governance.