The interesting number was not how much code Caylent could move. It was how much code could disappear. Arborgold, a software business in Teamfront’s portfolio, had five SQL Server clusters and roughly 2,500 stored procedures. A migration looked like a formidable translation job. Then Caylent’s tooling identified 1,850 procedures that were no longer used. The packing list had mistaken the attic for the household.
- Caylent builds, modernizes, and runs systems on Amazon Web Services.
- Its AI-assisted engineering combines automation with human review.
- The buyer’s question: does a smaller workload make the migration worth doing?
That distinction is a useful introduction to Caylent. Plenty of businesses want newer technology. Fewer want to pay for the archaeology required to get there. Old databases contain old assumptions. Applications carry deployment habits from an earlier organization. Somewhere, a team is keeping the whole arrangement alive while trying to build its replacement.
The code that stayed behind
Caylent’s Teamfront account puts the migration at approximately 220 engineering hours against an estimated 2,600 for manual work. Recurring database savings of about 77% were projected. Those are different measures: labor avoided is not the same as a smaller monthly bill, and neither tells you the consulting fee.
Most of the code didn’t need a ticket.
The technical destination was Babelfish for Amazon Aurora PostgreSQL, which preserves familiar T-SQL workflows. Routine translations were automated; complicated procedures still needed engineers. Regression testing checked the result. The practical lesson is refreshingly unglamorous: establish what runs, remove what does not, and test the surviving behavior before celebrating the new address.
An earlier engineering write-up describes the economic obstacle. Manual conversion looked too expensive and slow to justify. Caylent worked from a replica, converting procedures, testing against source data, correcting errors, and repeating. Automation changed the amount of work a migration required. That made a previously difficult investment easier to consider.
“We became our own first customer before we built anything for anyone else.”Caylent team, speaking to Anthropic
In Anthropic’s account, Caylent describes internal experimentation before customer deployment. Its agent-based system separates tasks such as assessment, translation, and validation. The revealing detail is the validation step: producing plausible code is only one part of the job. Someone must establish that the translated system still does what the business expects.
A holiday for Botox, a stress test for software
Another customer supplies a less abstract problem. On National Botox Day, unexpectedly heavy traffic burdened Allergan Data Labs’ application and hurt the digital customer experience. A promotional success had become an infrastructure problem. The company was already adopting serverless technology when it brought Caylent in to help with scalability, reliability, and costs.
Caylent helped establish container workflows, Kubernetes infrastructure, cloud foundations, and deployment automation. According to the customer story, deployment times fell from hours to minutes as monolithic pipelines gave way to smaller microservices. The team gained confidence to add workloads. Faster releases mattered because the people shipping software could attempt more without rebuilding their delivery machinery each time.
This is Caylent’s territory: the space between access to AWS and an organization’s ability to use it. Its services cover cloud foundations, migrations, application development, data platforms, security, and ongoing operations. Customers range from startups to Fortune 500 enterprises. The case-study library includes airlines, financial businesses, healthcare organizations, and software companies. The recurring problem is turning available technology into something a particular team can run.
One cloud, more of the work
Caylent began in 2015 with developer-tool roots. Its early fundraising profile describes container deployment software and monthly SaaS pricing, with JP La Torre as CEO and Stefan Thorpe as CTO. Today’s services business sells engineering judgment and delivery capacity alongside proprietary tools. The expertise has become part of what the customer buys.
The attraction is visible in Onyx CenterSource’s infrastructure consolidation. Growth through acquisitions had left the payments business with fragmented environments. Caylent migrated 160 virtual machines and 196 terabytes of data across European and US data centers, using repeatable infrastructure deployment and region-specific recovery planning. For a business assembled through deals, a shared operating environment is a practical benefit: fewer separate systems for the support team to navigate.
The company’s AWS focus gives it a recognizable position among broader consultancies and other cloud specialists. Slalom, Mission, or an internal engineering team may be alternatives depending on the assignment. Buyers choosing Caylent are choosing depth in a particular ecosystem. For organizations still deciding between cloud providers, that platform decision deserves attention before a delivery engagement begins.
AWS’s 2025 partner awards named Caylent the Consulting winner for Migration and GenAI. Those awards establish recognition within the partner network; a buyer still needs to match the team’s experience to the actual workload. A reference architecture cannot answer every question about an inherited application.
The expansion has been deliberate. East Los Capital invested $16 million in 2021. Gryphon Investors announced a majority investment in 2022. Caylent acquired Trek10 in October 2025, adding managed-services capabilities, and Pronetx in April 2026, extending its Amazon Connect customer-experience work. Valerie Henderson became CEO that March. Each move brings another part of the customer’s operating life within reach.
April also brought Anthropic Consulting & Engineering, a dedicated Claude practice. Caylent says the unit uses outcome-based pricing and fixed fees where measurement is feasible. Its broader culture emphasizes co-delivery and knowledge transfer; the careers page describes a remote organization operating that way since 2015. The customer team is meant to gain capability during the engagement.
The bill after moving day
The June 2026 operations launch makes the commercial proposition concrete. Agentic cloud operations packages start at $7,500 per month. A custom AIOps Blueprint starts at $125,000. These are starting prices for specific offers. A sensible comparison includes service scope, AWS consumption, implementation effort, and the customer’s own staffing needs.
Copy the method before copying the architecture. Inventory the estate. Put numbers against the manual alternative. Decide what evidence would prove the new system works. Keep engineers available for exceptions. AI-assisted migration becomes less persuasive when dependencies are poorly understood or expected behavior cannot be tested. Arborgold’s discarded procedures offer a modest place to begin: before asking how to move everything faster, check how much still needs to move.