The headline on 28 September was “erad Raises $22mln Series A led by MEVP to accelerate SME financing across the GCC”. Most of the coverage stopped at the number and the investor list. I read the release twice for a different reason: erad already lends in the UAE, it has said it will grow its technology team, and buried in the boilerplate is an operational claim that tells you precisely what kind of engineer the company has to hire next. By the end of that afternoon I had rewritten three Dubai job specifications we were about to publish.
What Actually Happened on 28 September
erad was founded in 2022 by Salem Abu-Hammour, Faris Yaghmour, Abdulmalik Almeheini and Youssef Said, and is headquartered in Riyadh. It provides Shariah-compliant working-capital financing of up to SAR 10 million to small and mid-sized businesses in Saudi Arabia and the UAE.
The round was led by MEVP, with new participation from 500 Global, Saudi Venture Capital, S60 Ventures, ANB Capital, Conjunction Capital and Araya Ventures, alongside existing investors Khwarizmi Ventures, Nuwa Capital, Aljazira Capital, Oraseya Capital and Joa Capital. Total equity raised now sits at roughly $32 million.
The operating numbers matter more than the round size: 8x year-on-year growth in Saudi Arabia, more than SAR 500 million (about $133 million) deployed cumulatively, and approvals delivered in an average of 48 hours. The stated use of funds is new financing products, further GCC expansion, and growth of the technology and commercial teams — with a specific push into industrial, logistics and manufacturing borrowers. Full details are in the Wamda report, the company announcement on Zawya and FinTech Global.
It was not the only Gulf round that day. Dubai-based Amaani raised a $5 million Series A led by BECO Capital in the same 24 hours. But erad is the one that changes hiring specifications, because erad is the one whose product is a piece of software making a credit decision.
Our expert take #1
“Approval in 48 hours” is almost never a machine-learning achievement, and reading it as one is how hiring goes wrong. The scoring model in a working-capital lender is usually modest — a few dozen features, a tree-based model, retrained monthly. The 48 hours is consumed by getting the inputs into a comparable shape: bank statements in a dozen layouts, VAT filings, accounting exports, settlement files from payment processors, trade licences, ownership records, and a stubborn share of documents that arrive as photographs of paper. When a lender compresses that to two days, they have solved ingestion and reconciliation, not inference. Hire accordingly.
The 48-Hour Claim, Read as an Architecture Diagram
If you want to know what a fintech is about to hire, take its most specific operational promise and work backwards to the systems that must exist for it to be true.
A 48-hour decision on facilities up to SAR 10 million implies a pipeline that is mostly automated but not fully. Somewhere in it there is a human credit officer, because at that ticket size no regulator in the Gulf and no sane risk committee lets a model disburse unattended. So the real engineering target is not “decide in 48 hours”. It is put a complete, reconciled, decision-ready file in front of a human in under a day, so the human has the remaining day to think.
That reframing changes the job. You are not hiring someone to raise AUC by two points. You are hiring someone who can shave hours off the slowest, ugliest, least glamorous part of the system.
The Shariah Constraint Is a Data-Model Constraint
The word “Shariah-compliant” in the release is not a marketing badge. It is a specification for your ledger.
A conventional loan accrues interest against an outstanding principal over time, and effectively every open-source lending schema, every off-the-shelf loan management system and every engineer who has worked at a conventional lender assumes exactly that shape. Shariah-compliant working capital is generally structured as a sale-based arrangement in which the return is a pre-agreed profit on an asset, agreed up front, not a rate accruing on a balance.
An engineer who models that as a loan with the field renamed produces a ledger that cannot be audited, cannot handle early settlement correctly, cannot be restructured without inventing numbers, and will not survive its first Shariah board review. I have watched exactly this happen at a Dubai lender that had to rebuild its core ledger eleven months after launch.
Our expert take #2
The scarcest fintech engineer in the Gulf is not an AI specialist. It is a backend engineer who has modelled Islamic finance instruments correctly at least once and can explain why the schema differs. That person is rare, they know they are rare, and they are almost never found by a keyword search for “fintech developer Dubai”. They are found by asking candidates to describe the data model they would use for early settlement — a question that takes four minutes and that a generic payments engineer cannot answer.
Hiring against the same talent pool this quarter?
We screen UAE fintech engineers on ledger modelling and ingestion architecture before they reach your calendar, so the first call is not spent discovering they have only ever worked on conventional loans.
Discutons-en — talk to our Dubai teamThe Three Job Specifications I Rewrote
Here is what changed, concretely, in specs we were publishing for Dubai clients that week.
1. From “Data Engineer” to “Document Ingestion Engineer”
The old spec asked for Python, Airflow, dbt and a warehouse. Every data engineer in the UAE matches that, so it sorted nobody. The new one asks for experience turning inconsistent third-party documents into a schema that a decision depends on, and it says out loud that a meaningful share of inputs are scans. The screening question is now: describe the worst input format you have had to parse in production, and what you did when it changed without warning.
2. From “Backend Engineer, Payments” to “Ledger Engineer”
Payments experience and lending experience are not the same skill, and conflating them is the most common specification error I see in the Gulf. Moving money is a distributed-systems problem. Owing money over time is an accounting problem. The rewritten spec screens on double-entry modelling, on early settlement and restructuring, and on the Shariah structure question above. Our guide to hiring fintech developers in DIFC covers the regulatory side of that profile in more detail.
3. From “ML Engineer” to “Risk Platform Engineer”
This is the one that saves the most money. The old spec was chasing a model-building profile at a model-building salary. The work, as the timeline above shows, is feature freshness, backtesting infrastructure, policy versioning and override auditing. That is platform engineering with a risk vocabulary, the candidate pool is far deeper, and the people in it are generally happier doing it. If genuine model work is also needed, our note on hiring AI and ML engineers in Dubai separates the two profiles properly.
The Detail Almost Everyone Missed: Industrial, Logistics, Manufacturing
erad said the new capital targets capital-intensive sectors — industrial, logistics and manufacturing borrowers. That single line implies a different data problem from the e-commerce and services SMEs that alternative lenders in the Gulf usually start with.
An online retailer hands you clean, high-frequency, machine-readable settlement data from a payment processor. A manufacturer hands you purchase orders, invoices with 60 to 120 day terms, inventory that is collateral, equipment with a resale value, and a receivables book concentrated in a handful of counterparties. Cash flow is lumpy and seasonal rather than daily.
So the features change, the fraud surface changes, and the engineering changes with them. Whoever builds this needs to be comfortable with counterparty concentration, with document-heavy evidence, and with the fact that the signal arrives monthly rather than hourly. That is a different hire from the one most Dubai fintechs made in 2024 and 2025.
Our expert take #3
A funded competitor is a compensation event before it is a competitive one. erad is not going to take your customers in Dubai next quarter. But it will be interviewing the same forty or so people in the GCC who have built regulated lending systems, and it now has the balance sheet to close them. If your plan was to hire that profile in November at the band you benchmarked in June, the band has moved. The cheapest defence is not money — it is a job specification that describes a genuinely interesting problem, because this particular cohort reliably chooses the harder system over the larger number.
What I Would Do This Week
- Re-benchmark before you post. If your UAE fintech salary bands are older than a quarter, they are wrong. Our Dubai salary negotiation guide has the current structure, including the allowances that move total package more than base does.
- Split payments experience from lending experience in every specification you have open. They are different jobs and you are currently attracting the wrong one.
- Move your hardest question to the first call. For lending, that is the early-settlement data model. Four minutes, and it sorts the market.
- Decide whether you are hiring for the model or the pipeline. The timeline above says pipeline, roughly 80% of the time, at a materially lower cost.
If you are building the product rather than the team, our walkthrough of what it takes to build a fintech app in the UAE covers the licensing and architecture sequence. And if the same profile is on your Singapore roadmap, the constraints differ enough to be worth reading separately — our colleagues cover hiring senior backend engineers in Singapore for the APAC side of the same build.
Competing for the same forty engineers?
We source fintech engineers across the UAE and pre-screen them on ledger modelling, ingestion architecture and regulatory exposure — and we will tell you when the role you have written is really two roles.
Discutons-en — brief our Dubai teamFrequently Asked Questions
What exactly did erad announce on 28 September 2026?
erad, a Riyadh-headquartered SME financing platform founded in 2022, announced a $22 million Series A led by MEVP, with participation from 500 Global, Saudi Venture Capital, S60 Ventures, ANB Capital, Conjunction Capital and Araya Ventures alongside existing backers including Khwarizmi Ventures, Nuwa Capital, Aljazira Capital, Oraseya Capital and Joa Capital. The round takes total equity funding to roughly $32 million. erad provides Shariah-compliant working-capital financing of up to SAR 10 million to small and mid-sized businesses in Saudi Arabia and the UAE, and says it uses data and AI to approve applications in an average of 48 hours. The company reported 8x year-on-year growth in Saudi Arabia and more than SAR 500 million deployed cumulatively, and said the money will fund new financing products, GCC expansion and growth of its technology and commercial teams.
Why does a Saudi funding round matter for hiring in Dubai?
Because erad already lends in the UAE and has explicitly said it will grow its technology team while expanding across the GCC. Engineers who can build regulated lending systems are a small, shared pool across Riyadh, Dubai and Abu Dhabi, and a funded competitor hiring into that pool moves the price for everyone in it. The practical effect for a Dubai fintech is not that erad will poach your team next week. It is that the profile you were planning to hire at a certain band in November is now being courted at a higher one, and your job specification needs to compete on problem quality rather than on salary alone.
Is a 48-hour credit approval a machine-learning problem or a data problem?
Overwhelmingly a data problem. The model that scores an SME is rarely the hard part, and in most working-capital lenders it is a gradient-boosted tree over a few dozen features that a competent data scientist can build in weeks. What consumes the 48 hours is acquiring, normalising and reconciling the inputs: bank statements in inconsistent formats, VAT filings, accounting-package exports, payment-processor settlement data, trade licence and ownership records, and in the Gulf a meaningful volume of documents that arrive as scans. Engineers who have shipped this understand that the latency budget lives in ingestion and reconciliation, not in inference. That distinction is the single most useful thing to screen for.
What does Shariah compliance change about the engineering job specification?
It changes the domain model rather than the technology stack. A conventional loan accrues interest against a principal over time, and most off-the-shelf lending schemas assume exactly that. Shariah-compliant working capital is typically structured as a sale-based or commodity-based arrangement in which the economics are expressed as a pre-agreed profit on an asset, not as interest on a balance. An engineer who models it as a loan with the word interest renamed will produce a ledger that cannot be audited, cannot be restructured correctly, and will fail its first Shariah board review. This is a domain-modelling skill, and it is why lenders in the Gulf consistently report that their hardest hire is a backend engineer with actual Islamic finance exposure rather than a generic payments engineer.
Panos Petropoulos
Web Development Expert at HireDeveloper.ae. Reviews engineering job specifications and technical screens for UAE fintech and platform teams.