We build AI that
grows your
business.

We turn your company’s expertise into AI products and software that improve how your business runs. Our team builds them with you and keeps improving them after launch.

Let’s talk
Explore what we can build

You’ve built
the advantage.
Let’s make more of it.

You bring what took years to build: the know-how, and the customers who trust it. We bring our own software and the senior engineers who know how to build on it.

Make your expertise
something people can use.

Turn what your company knows into a product your customers will pay for. We start with who it serves and whether the economics work, then build what deserves to exist.

Grow without every cost
growing with you.

Automate the repetitive work and manual handoffs that slow your operations down, so quality goes up and your people spend their time where it matters.

Raise the standard
across your whole team.

Give your people a place to practice the conversations that matter, with feedback built on your own playbook. We involve them early, so the change sticks.

We assess the value and requirements with your team before deciding what to build.

We partner with selected clients
to future-proof their company.

Software we’ve already built,
so you don’t start from zero.

We bring our own software into each project and adapt it to your knowledge and your people.

Atlas / Caliber

Practice difficult conversations with scenarios from your business and feedback shaped around your team’s playbook.

Caliber workspaceIn development
Objection handlingSwitching providers
Your scenario

Explore the concern. Understand the impact.
Respond using your playbook.

Built on your knowledge, held to your standards.

Content Engine

Turn expert interviews and approved knowledge into content, with your team shaping the voice and approving publication. We use it ourselves every week.

Content workspaceSimplified for illustration

Knowledge

Interview
Questions

Drafts

Article
Email
Social post

Review

Facts
Voice
Approval

Publication stays with your team.

From expertise to publication.

More FP tools…

Caliber and Content Engine are two of the tools we bring into projects. Tell us what your team needs and we’ll show you what else we’ve built.

Talk with us

Start with the
opportunity
that matters most.

30
days to begin execution.

Launch timing depends on scope, system access, testing, and approvals.

Find what’s
worth building.

We start with your business and the people it serves, and agree on what success looks like before anything is built.

Build it inside
your systems.

Our senior engineers work next to the experts whose knowledge the product depends on.

Release it,
then improve it.

We release in stages, with monitoring and a way to pause, and measure progress against what we agreed on in step one.

Each month we contact
more than 2,000 engineers.
We accept 12.

Every engineer on our projects goes through a five-stage process. The last stage is a technical conversation with Jess Mah, who personally clears every engineer we accept.

  1. 2,000+contacted each month
  2. ~250screened
  3. ~30advanced
  4. 12accepted

How we assess them.

HackerEval combines practical technical work with a consistent assessment of coding, AI fluency, prompting, architecture, product thinking and communication.

Our engineering standard
  1. 01 Initial screen
  2. 02 Technical assessment
  3. 03 EQ & behavioral interview
  4. 04 Paired AI challenge
  5. 05 Final interview with Jess Mah

Prior work by engineers on our team.

Anonymized earlier engagements.
Healthcare AIA citation-backed RAG over
4,000 clinical guidelines.Senior engineer
Brazil · 7 yearsInsurtechA production RAG pipeline,
built in three weeks.Senior backend
Colombia · 8 yearsAI-native operationsMulti-agent orchestration
designed for scale.Senior backend
Brazil · 6 yearsFintechAI checkout experiences,
shipped in production.Full-stack
Argentina · 7 years

The healthcare build reached its first production deploy in 11 days, scoring 87% on its evaluation suite against an 80% target. That was faster than usual: comparable builds reach production in four to six weeks. These are prior delivery examples, not forecasts for your project. Names are shared under NDA.

Engineers fluent
in your tools.

Coding agents, language models, automation, creative tools and cloud infrastructure: the stack your product will run on is the one our engineers already work in every day.

Meet the engineers

Long after
launch.

Launch is the beginning
of what we can build together.

We help your people use what we built, fix what needs attention and keep looking for the next thing worth building as your business evolves.

The conviction behind our work

Your guardrails
apply from day one.

We involve your technology and legal teams from the start. Clear boundaries and shared responsibilities make room for confident progress.

Data boundaries

Agree on what information is needed, who can access it, where it can be processed, and which provider terms apply.

Quality and human oversight

Define relevant test scenarios, acceptable performance, and where people review or intervene.

A controlled release

Set deployment stages, approval responsibilities, monitoring, and ways to pause changes.

Clear ownership and support

Agree on who operates the capability, handles incidents, makes improvements, and supports it after launch.

We’ve built companies.
We know what’s at stake.

Meet our co-founders. They’ve built businesses, led teams, and lived with the decisions that follow a launch.

Jess Mah

Co-founder of inDinero and Executive Chair of Mahway. Company-building experience and a background in computer science and data science. She runs every final technical interview.

Profile

Andrea Barrica

Co-founder of inDinero, founder of O.school, and former venture partner at 500 Startups. Experience building companies and helping founders grow.

Profile

Gabe Murillo

Founder of EnjoyTalent. More than 15 years in senior talent acquisition, including building the sourcing strategy at Shopify Logistics.

Profile

Notes
from the field.

What we learn building AI products, and how we choose the people who build them.

All insights
WorkflowWe Shipped a Production RAG Pipeline with Claude Code in 11 Days — Engineer-Level Velocity DataApr 30, 2026 · 10 min readResearchWhy I'm Betting on LATAM AI Engineers — A Note from Jess MahMay 6, 2026 · 7 min readVettingWhat Jess Mah Looks for in a Senior AI Engineer — The 5-Minute FilterMay 6, 2026 · 7 min read

Guides for teams building with AI.

Decision frameworks, tool comparisons and the research behind how AI-native teams build, ship and decide what to build next.

All resources
Most read

Questions,
answered.

Have a different question?
Do we need to know exactly what to build?

No. Bring your ambition and your knowledge of the business. We’ll work with your leaders to assess the opportunity, the economics, and what makes sense to build first.

How does the partnership work?

We partner with a small number of established companies at a time. We work with your leadership on where AI can create the most value, build what we agree on with your experts and keep improving it after launch. The scope and the way we work together are agreed before anything starts.

Who does the work?

Our own senior engineers, alongside our AI strategists and product leaders, working inside your systems with your experts. You don’t hire or manage anyone.

How are your engineers evaluated?

Each month we contact more than 2,000 engineers and accept 12. Our five-stage process includes a technical assessment, a behavioral interview, a paired AI challenge and a final interview with Jess Mah. HackerEval supports practical evaluation across six dimensions, from coding and architecture to product thinking and communication.

Who owns what we build?

You do. What we build for your business is yours, and so is your data. Access controls, provider terms and release approvals are agreed with your technology and legal teams.

Will our data train an AI model?

Using company knowledge does not automatically mean training a model. We agree on access, retention, provider terms, and any proposed training with your organization.

Do you still place engineers?

No. We don’t place or staff engineers. Our engineers work on the projects we build with our partners.

What happens after launch?

We keep working with you: helping your people use the software, reviewing performance, fixing what needs work and deciding what to build next. The operating model and ongoing responsibilities are agreed with you.

How much bigger
could this
become?

Tell us what you have in mind. If it’s a fit, we’ll show you how we’d build it. If it isn’t, we’ll say so.
Every proposal is read by one of our founders.

Let’s talk
Open the film

Our products

    Discuss it with us