I have now spent 10 years at Cheesecake Labs. The company I joined was born in the mobile era, and I watched new platforms change what our clients could build. I have never seen a moment with more potential than this one.
What excites me the most is not (only) access to another frontier model. It’s the insight and efficiency companies gain when production AI connects their systems, data, products, and people.
That is why Cheesecake Labs is evolving. We are bringing 13 years of production engineering into a sharper proposition. Modernize the systems beneath the business, and make data trustworthy and usable. Then put AI agents into production, with the controls and engineering discipline required to make them work.
The opportunity is implementation
RSM found that 91% of middle-market respondents use generative AI. However, only 25% of users say it is fully integrated across core operations. Seventy percent said they needed outside help to get more from it. BCG found that only 6% of the companies in its analysis qualified as AI leaders. Their three-year, industry-adjusted shareholder returns were 9.3 percentage points above the sample median.
OpenAI and Anthropic have both announced hands-on deployment businesses centered on engineers working alongside clients. Their direction reinforces what we see in the mid-market and regulated organizations we have served for 13 years. Model access is only part of the work.
The difference is implementation: connecting AI to proprietary context, redesigning a real workflow, and operating it safely at scale before competitors do.
A natural next chapter for Cheesecake Labs
We have shipped more than 300 digital products, modernized aging platforms, connected fragmented systems, built data foundations, designed regulated products, and embedded senior engineers inside client teams.
Now we’ve joined Anthropic’s Claude Partner Network and are a consulting partner of Databricks, Snowflake, and dbt.
AI gives that experience more leverage. It can accelerate code, tests, documentation, and data pipelines. Agentic delivery can help teams map legacy systems, document dependencies, build test coverage, and migrate through smaller, validated releases.
In one current modernization engagement, we previously would have scoped the work at 18 to 24 months. Instead, it is being delivered across only 5 months by a smaller, agent-augmented team, through validated production releases.
Faster engineering alone does not change a business. Results come from connecting what gets built to live systems, governed data, real workflows, and the people responsible for them. Production AI needs evaluations before launch, guardrails and human escalation in production, and continuous monitoring after release.
The context that makes AI useful lives in the records and rules behind actual decisions: support history, eligibility criteria, product configurations, inventory, pricing, CRM and billing relationships, and the outcomes of past actions.
As frontier models become more accessible, the advantage shifts to companies that keep this context current, governed, and available at the point of decision.
What we help clients do now
Four capabilities define what we build:
AI Implementation: Agentic workflows in production, grounded in trusted data, with governance and controls built in from the start.
Application Modernization: AI-assisted discovery, documentation, testing, and phased migration that modernize aging systems while keeping core operations running.
Data & Analytics: Modern data platforms, reliable pipelines, semantic models, lineage, quality controls, and permissions that analytics and AI can trust.
Product Engineering: Strategy, design, and engineering across mobile, web, IoT, blockchain, cloud, and LLM and RAG integrations, from first release through scale.
Two capabilities define how we deliver. AI-Accelerated Delivery applies agents across discovery, build, review, testing, and release while keeping production standards intact. Embedded Engineering Teams bring senior engineers into the client environment, nearshore, on site, or US-based. They work alongside internal teams or own delivery end to end.
Together, they guide our clients through every stage: assessing what matters, modernizing systems and data, implementing AI, and scaling what proves its value.
What production AI looks like in practice
In regulated environments, a practical sequence starts with visibility and governance, then builds the data and AI architecture. Only then do we introduce narrowly scoped agents, with defined permissions, evaluations, human approval, and audit trails.
Forward-deployed engineers make this practical. They work inside the client’s systems, tools, and workflows from assessment through launch. Frontier AI labs now describe this model as forward-deployed engineering. We have worked inside client teams this way for 13 years.
Our public work shows the same foundations in production. For example, with SWIRL, we helped connect enterprise AI search to more than 100 platforms. SWIRL clients report up to 80% less time spent searching internal data. With Knapsack, we helped build a privacy-first AI workflow platform. The platform carries SOC 2 Type I certification, HIPAA compliance, and GDPR-aligned practices.
A new brand for the company we have become
Our new visual and verbal identity makes this evolution visible. Cheesecake Labs is now an AI services and engineering company for clients that need modern systems, trusted data, products built to scale, and senior engineers who can take AI into production.
This is not a departure from our product engineering history. It is the next expression of it.
After 10 years here, that is what excites me most: helping established companies modernize faster, make their business context usable, and put AI to work in the workflows that matter.
Whether you are looking to modernize a legacy application or take AI agents into production on a trusted data foundation, let’s talk.
