115
tests across six layers
none skipped
Baghdad, Iraq · Tech Lead, Vodex IT Solutions
Distributed sync engines, multi-tenant SaaS, real-time fleet platforms, IoT, payment rails. I scope them directly with clients and coordinate delivery across teams in six countries. I load-test them until they break, publish the numbers, and say what is still not fixed.
Trace · one request, end to end
Delivery drivers work for hours with no connectivity, then sync. The hard problem is not the sync, it is what happens when the field and the office disagree.
PRJ-001 · platform · 2026 · source public · theeCaesar/Sanad
One Postgres transaction per mutation covering six writes across five tables. Events go to an outbox inside that transaction, so a broker outage becomes a delay rather than a data-loss event. Sockets are a view, never a write path.
PRJ-002 · platform · 2025
Keeps selling when the internet drops, then reconciles every till operation exactly once. Operations are idempotent by client-generated id and stock is guarded at the database, not in application logic.
PRJ-003 · platform · 2025 · under NDA
Three protocols, one process. The locks speak a proprietary binary protocol over raw TCP, riders speak HTTP, the operator dashboard speaks WebSocket, all terminating against a single domain layer. Two riders scanning the same QR is the defining race, and a compare-and-swap makes the guard and the claim the same operation.
Verify it yourself
PRJ-001's source is public: the conflict resolver traced above, its 48 tests, and the load-test write-up are all in the repository. The other twenty-four systems are client-owned, so they're described rather than shown.
Load tested to failure, then fixed
A 300-device reconnect storm broke it at 44.66% sync failures. Pool exhaustion that never threw an error, because the driver's timeout bounds the TCP connect and not the wait for a pooled client. Rebalanced to 20.15%, with throughput up 38%. p95 is still pinned, and that is not done.
115
tests across six layers
none skipped
0.7s
48 resolver tests
pure functions, no I/O
7
decisions written up
each with its tradeoff
3
transports on one core
REST · gRPC · GraphQL
Terracotta marks where I'm strongest
Languages
Frameworks
Data
Infrastructure
Security & networks
Practice
Certified: eJPT · CCNA-level networking · Cisco course assistant

Mohammed Hussein
Baghdad, Iraq · GMT+3
I run discovery with the client, design the architecture, own the infrastructure it runs on, and coordinate the team that builds it. Transport choices, database choices and consistency models are decisions I make and then bring the team through, including reviewing the code that comes out the other side.
Open to
Senior or lead backend roles, remote anywhere or on-site in Baghdad, and consulting engagements at project inception.
Most delivery risk is decided at inception, on scope, feasibility and sequencing. That is where I am most useful, and it is where I prefer to come in.
01 · Discovery and scoping
Requirements workshops with technical and non-technical stakeholders, turned into specifications engineers can build from, plus the architecture and service boundaries to match.
02 · Architecture review
Reading an existing system and naming what will break, including inherited codebases. I took over an eighteen-model TypeScript and SQL platform, migrated the stack, and delivered it at sixteen models with full endpoint documentation.
03 · Build lead
Owning delivery end to end: infrastructure, release and rollback, secrets, domains and TLS, and coordinating the team and any external vendor teams through to production.
I have directed vendor engineering teams and hardware manufacturers across Canada, the United States, India, China, Egypt, Algeria and Saudi Arabia, across time zones and engineering cultures.