You are shaping an early SaaS product.
Turn the product idea into one complete user journey and a first release the team can learn from.
CUSTOM SOFTWARE, AI & PRODUCT ENGINEERING
Custom software for complex work.
We design and build custom applications, practical AI, and developer tools for workflows that off-the-shelf software cannot support.
Product engineering · Systems thinking · Careful delivery
WHEN THE WORK NEEDS SOFTWARE
A useful fit for teams at a turning point.
The project may be a new product, a system that has outgrown its first version, or a repeated workflow that deserves a clearer path.
Important work is split across tools.
Make the handoffs, records, and exceptions easier to follow with a focused internal system or integration.
A technical workflow needs a better interface.
Give developers or model-assisted processes clearer contracts, useful feedback, and safer failure paths.
WHAT WE BUILD
All capabilitiesSoftware shaped around the real problem.
From customer products to internal systems, we focus on making the underlying workflow clearer, more useful, and easier to maintain.
Custom software
Purpose-built applications for a problem off-the-shelf tools cannot solve.
SaaS product development
From early product shape through the first reliable release and beyond.
AI applications and agents
Useful model-powered features with controlled inputs, tools, and failure paths.
Developer tools
Interfaces, APIs, and utilities that make technical work easier to understand and repeat.
Workflow automation
Replace brittle handoffs with observable, recoverable software workflows.
SELECTED WORK
All workSoftware in the real world.
Explore AI products, research workflows, and engineering tools through the work itself.
ForgeLoop
A delivery system connecting GitHub issues, bounded agent tasks, self-hosted runners, verification evidence, and pull requests.
Explore the project
Witnessrail
A research engine and interactive showcase exploring guarded browser actions and independent verification for payer enrollment.
Explore the projectHOW WE WORK
A clear path from question to software.
Keep the problem visible at each step. Make decisions with the people who will use and maintain what gets built.
Understand the work
We get specific about the people, constraints, existing systems, and decision the software needs to support.
Shape the system
We agree on a useful first scope and make the product, data, and technical choices visible before they become expensive.
Build and refine
We deliver working software in clear increments, then use feedback and real use to guide the next step.
THE PEOPLE BEHIND IT
A small studio, close to the work.
Jared Hooker and Esraa Kamel bring experience across full-stack product engineering, Python backend systems, and production AI workflows.
01Co-founder
Jared Hooker
A full-stack software engineer who builds and operates SaaS products, AI applications, and developer tools, from architecture through production.
Founder profile
02Co-founder
Esraa Kamel
A Python backend and agentic AI engineer who builds production APIs, retrieval systems, and durable workflows with explicit controls and evaluation.
Founder profileWRITING & NOTES
All writingIdeas from the build.
A Business Dashboard Should Tell You What Needs Attention
How to design a business dashboard around decisions, meaningful comparisons, trustworthy data, and a clear route from a metric to the work behind it.
By Esraa Kamel · 4 min readRead
Design SaaS Permissions Before Your First Enterprise Customer
A practical guide to SaaS access control: define who can act on which records, separate roles from account boundaries, and test revocation and denied access.
By Jared Hooker · 4 min readRead
Before You Build Custom Software, Map the Work First
A practical discovery guide for teams weighing custom software: trace the real task, its handoffs, decisions, exceptions, and evidence before writing a feature list.
By Esraa Kamel · 4 min readRead
Browse more notes in the studio notebook.
A FIRST CONVERSATION
You can start before the spec is finished.
A useful first discussion is about the people, the task, what happens today, and which constraints already matter. From there, we can decide whether a small discovery step, a technical review, or a build is the right next move.
What should I bring to the first conversation?
A rough description of the task, who needs to do it, what makes it difficult today, and any deadlines or system constraints you already know. A complete specification is not required.
Can you work with an existing product or codebase?
Yes. The starting point is to understand its current users, dependencies, and maintenance needs. That helps distinguish a focused improvement from a larger replacement.
Do you build AI into every product?
No. Use a model when flexible interpretation adds value and its outputs can be evaluated and bounded. Stable rules, permissions, and important decisions should stay explicit.
HAVE A PROJECT IN MIND?
Start a conversation