We are dataops experts
We can help you modernize your data operations so you can serve your customers better and faster
Modern data systems are mostly boring-but-proven technologies, while AI and machine learning are super interesting but still emergent. We like both, and are good at both — we can help you modernize and improve your real work, but also set you up for cool experiments with emerging technologies like large-language models
Data Pipelines
ETL/ELT, Dbt, Airflow, Dagster
This is the most important, and often messiest, step in any data program, and we are fast and good at it. We can also introduce modern software development practices so that your staff can properly test pipelines before they go live in production.
Make Old Systems New!
Lift and Shift
We can help you move your data workflows to public or private cloud tech stacks. We like old systems because they contain well-understood, valuable insights, and if you move them to cloud delivery models it gives you a great platform to accelerate your data program.
Data Warehouse
Snowflake, Redshift, S3, BigTable
We have experience with big data platforms and can help build, optimize or manage a big data platform so that it securely and seamlessly integrates with analysts' tools and workflows.
Data Analytics
Staff Augmentation
Our experienced, senior data analysts can take on projects so you can accelerate your program. We have experience with data operations in both large and small companies.
Analysis Delivery
Get Results in Users Hands
We understand how to deliver on established platforms like Tableau, Microsoft 365 Graph API (Excel, PowerBI), AirTable, JasperReports and BIRT. We can also formulate a modern, real-time approach to results delivery.
Machine Learning
Spark/Databricks, SkyPilot
We can help you build a scalable machine learning platform that doesn't break the bank. We build the infrastructure so your analysts can easily deploy highly parallelized algorithms like monte carlo simulations, k-means clustering, k-nearest neighbors and more.
Large Language Models
Embedded Code Completion
We can integrate open LLMs such as Llama2-Code-Llama to help your data analysts generate code without anything ever leaving your cloud.
Metrics, Observability and Trace
Up Your Game for Real-Time Production
Complex, production-facing systems indicate problems long before failure. Because we've done this for so long we have cool ideas around materialization that strike a good balance between the needs of production-debug and trend analysis without costing a fortune. (Logging and metrics often are a top cloud account expense.)
Production Signals
APIs that provide real-time production signals
We can help you build APIs that deliver real-time analytics back to production systems that meet performance and security requirements. Have an idea for a fraud signal? We can help you integrate that into production.
GitOps
Automate Everything
With GitOps your analysts simply commit their work to a git branch and it deploys to either dev or prod depending on the branch. Now you have total auditability of what is running and who deployed it. Rollback is simply a git revert, and you can rebuild an entire environment with one command (or replicate an environment for disruptive development testing).
Cloud Debug
Debug pipelines from a Laptop
We can provide a way for analysts to stop on a breakpoint in their local Pycharm against an Airflow DAG running in the cloud. This is incredibly transformative, because it's almost impossible to recreate data pipeline conditions in test, and the weirdest stuff seems to always happen in the pipelines. You get to more robust pipelines faster.
Security
IAM & Role Management
Your data contains PII and trade secrets, but often analysts download it to their laptops because the data tools setup makes it hard to work with the data safely. We can fix this so the analyst can work in a way that is not only safer, but much easier.