Application re-architecture
Monolith to microservices, one capability at a time, with the strangler fig approach.
Big-bang rewrites of large systems are risky. Birbal AI breaks monoliths apart gradually: we map the domains, put a routing layer in front, carve out one capability at a time as a service with its own data, and retire that part of the monolith once the service is proven. Sometimes the better answer is a well-structured modular monolith, and we'll tell you when.
- Strangler fig approach
- Domains mapped first
- Data split carefully
- Fixed-price or team-based
Monolith, taken apart gradually
No big-bang rewriteAuthorized vendor to
- Ministry of Finance
- Ministry of Interior
- DGE Abu Dhabi
- Etihad Aviation
- EDGE Group
Delivered for
- Hindustan Unilever
- Nokia
- Linde
- USA Today
- JSL
Procurement-ready
- Emirati national company
- National ICV certificate
- Abu Dhabi Government Procurement Gate
- UAE Ministry of Finance, federal procurement
- Registered ADNOC vendor
What we deliver
Re-architecture, from domain map to retired monolith
Domain mapping
Capabilities, data ownership and dependencies mapped from the code and from the people who run the business.
Target architecture
Service boundaries, APIs, events and data ownership designed, including what should stay in the monolith.
Strangler fig migration
A routing layer in front of the monolith sends each capability to its new service as it is ready.
Data separation
Shared tables split with care, through replication, events or APIs, so each service owns its data.
Platform
Containers, Kubernetes or managed services, CI/CD, observability and service-to-service security.
Modular monolith
When microservices aren't worth it: the monolith restructured into clear modules with enforced boundaries.
How we work
How a re-architecture runs
- 01
Map
We trace calls, shared tables and teams, and draw the domains and their dependencies.
- 02
Design
Service boundaries, APIs, events and data ownership are agreed, with an order of extraction.
- 03
Put the facade in place
A routing layer goes in front of the monolith, so traffic can move capability by capability.
- 04
Extract and prove
Each capability is built as a service, tested against the monolith's behaviour and switched over.
- 05
Retire
The matching part of the monolith is removed once its service is proven in production.
The approach
Why gradual, and why not always microservices
Why gradual
Martin Fowler's Strangler Fig article describes replacing a legacy system step by step, growing the new system around the old one until the old one can be switched off, instead of replacing it all at once. Each step is small enough to test and reverse.
Why not always microservices
Each service adds deployment, monitoring and data work. Where teams and scale don't need it, a modular monolith gives most of the benefit for less.
Where AI helps
Where AI helps in a re-architecture
AI maps and drafts; architects decide.
Mapping dependencies
Calls, shared tables and hidden couplings traced across the codebase.
Suggesting boundaries
Candidate service boundaries drafted from the code and the data, for architects to decide.
Drafting extracted services
First drafts of service code and APIs from the monolith's logic, reviewed by engineers.
Writing contract tests
Tests that pin each API's behaviour, so services and the monolith stay in step.
What we work with
What we work with
Monoliths
Java, .NET, PHP, Python and Node.js applications.
Services
Spring Boot, ASP.NET Core, Node.js and Python services.
Integration
REST and gRPC APIs, events and message queues.
Platform
Docker and Kubernetes, Azure Container Apps, or Amazon ECS and EKS.
Observability
Logs, traces and metrics across services.
Delivery
A CI/CD pipeline for every service.
How to work with us
A delivered project, or specialists on your team
| The question | Fixed-price project | Specialists on your team | Dedicated development center |
|---|---|---|---|
| Who manages the work | We do, to an agreed scope | You do, day to day | You set priorities; we run the team |
| What you pay for | A fixed price for the scope | A monthly rate per specialist | A monthly rate per engineer |
| Best when | You want capabilities extracted to an agreed scope | You have teams and need architecture skills | You want a long-term platform team |
| Where they work | At our center in Abu Dhabi | At your office in the UAE, or remotely from the UAE | At our center in Abu Dhabi |
Pricing
Rates for specialists on your team
For team-based work: one monthly fee per specialist, with no recruitment fees. Fixed-price projects are quoted after discovery.
Onsite in the UAE
- Junior to mid-level
- AED 20,000–35,000
- Senior and specialized
- AED 35,000–45,000
Per month. Includes local benefits, visa processing and onsite support.
Remote
- Junior to mid-level
- USD 3,000–5,000
- Senior and specialized
- USD 5,000–8,000
Per month. Engineers work remotely from the UAE. Includes project management and quality assurance.
Final pricing depends on expertise, project duration and requirements. We send a detailed proposal once we understand your needs.
Who it is for
When a monolith needs breaking up
Releases that take weeks
Every change needs a full regression test and a coordinated release.
Teams that block each other
Many teams working in one codebase and one deployment.
Parts that need to scale separately
One busy feature forcing the whole system onto bigger servers.
A new front end on an old core
Apps and partners that need clean APIs the monolith can't offer.
Specialists you can add
Hire the specialists behind this work
Track record
Trusted by UAE government entities and global brands.
Authorized vendor to five UAE government entities, with projects delivered for Hindustan Unilever, USA Today, Nokia and Linde.
- Startups and enterprises served by our engineering teams
- 150+
- Startups and enterprises served by our engineering teams
- Average rating on Clutch, Trustpilot and Glassdoor
- 4.9/5
- Average rating on Clutch, Trustpilot and Glassdoor
- Client satisfaction rate
- 98%
- Client satisfaction rate
- UAE government entities we are an authorized vendor to
- 5
- UAE government entities we are an authorized vendor to
Hear it from our clients
Client testimonial
E-commerceClient testimonial
HealthcareClient testimonial
E-commerceClient testimonial
E-commerce
Why Birbal AI
A team you can hold to account
An Emirati national company
ICV certified, with our office and development center in Al Zahiya, Abu Dhabi.
Engineers based in the UAE
Every engineer is based in the UAE and can work onsite at your premises in the UAE, or remotely.
Your environment, your rules
We work inside your environment and follow your data rules, with NDAs as standard.
You own the work
Code, configurations and documentation are transferred to you on delivery.
Sources
Where the outside facts on this page come from
Free estimate
Get a free re-architecture estimate in 48 hours
Share your requirements, an inventory or an RFP. You get a scope, timeline and cost range within 48 hours, then a fixed-price proposal after a short discovery.
- Free and no obligation
- Fixed-price or team-based options
- NDA before we see any code
4.9/5 average rating · 98% client satisfaction · 150+ clients served
FAQ
Questions buyers ask
What is the strangler fig pattern?
A way to replace a legacy system gradually: new services take over one capability at a time, behind a routing layer, until the old system can be switched off. Martin Fowler named it after strangler fig vines.
Do we need microservices?
Not always. If a few teams work on the system and it scales well enough, a modular monolith may be the better choice. We recommend one after mapping your system.
What happens to the shared database?
It is split gradually. Each extracted service gets the data it owns, kept in sync with the monolith through replication, events or APIs until the old tables can be retired.
Will users notice the move?
They shouldn't. Traffic moves capability by capability behind the routing layer, and each step can be reversed.
How long does it take?
It depends on the size of the monolith and how many capabilities you move. The first service usually goes live early, and the rest follow in a planned order.
How much does a monolith to microservices migration cost?
It depends on the size of the code, the number of interfaces and how much test data exists. You get a scope, timeline and cost range within 48 hours of sharing your requirements, then a fixed-price proposal after a short discovery. Engineers for your team start from USD 3,000 a month remotely or AED 20,000 a month onsite in the UAE.
Where does our code go when you use AI on it?
Only where you agree. The AI tools can run inside your own environment with a private model, the approach behind our Mustashar product, or use an AI service your organization has already approved. We sign an NDA before we see any code, and engineers review every change the AI drafts.
Who owns the converted code and documentation?
You do. We sign NDAs as standard and transfer the code, tests, configurations, documentation and IP on delivery.
Plan how to take your monolith apart
Share your system, its size and your goals, and get a scope, timeline and cost range in 48 hours.
Get my estimate