Move data without moving problems.
Batch and event-driven pipelines designed for recovery, observability, and schema changes—not just the happy path.
Data engineer · Sana'a, Yemen
I build the systems
between noise
and knowing.
Reliable pipelines, thoughtful data models, and platforms that give teams answers they can trust.
12M+
Daily records
99.98%
Pipeline SLA
0.00
Variance
Available for data engineering opportunities
Scroll to inspect
What I build
Data systems that stay useful after launch.
I focus on the parts of data work that compound: dependable movement, clear models, and visible system health.
Shape raw records into shared meaning.
Warehouse models and semantic layers that make metrics consistent, discoverable, and immediately trusted by product and finance.
Make the invisible visible.
Automated tests, anomaly detection, and operational alerts so silent errors get caught before they become executive decisions.
Teach systems to reason, not just run.
Agentic workflows that inspect data drift, invoke tools, and auto-heal pipeline anomalies—anchored on strict verification contracts.
- 6h reduced to minutes
- to detect → now minutes
- 0
- unexplained variance
- 24/7
- agent-monitored
A real problem, untangled
When the daily revenue
number stopped matching.
-
Problem
Finance flagged a recurring gap between the dashboard total and the bank statement—small enough to ignore, persistent enough to erode trust in every other number downstream.
-
Approach
Rebuilt the reconciliation step as an agentic check: it plans a trace across orders, refunds, and payouts, calls the warehouse directly, and only escalates when it can name the cause.
-
Result
The same class of mismatch now gets caught and explained before anyone has to ask about it—turning a monthly fire drill into a quiet log line.
How I think
Build for the question.
Engineer for everything after it.
A good data system starts with the decision someone needs to make, then works backward to the source.
-
01
Trace the decision
Start with who needs the data, what they need to decide, and how fresh the answer must be.
-
02
Design the path
Choose the simplest reliable architecture, with explicit ownership and failure behavior.
-
03
Prove the signal
Validate quality at the boundaries and make health legible to the people operating the system.
-
04
Leave it clearer
Document the reasoning, reduce hidden knowledge, and make the next change safer than the last.
Behind the systems
Curious by default.
Deliberate by practice.
I’m Mohammed Babaqi, a data engineer who enjoys turning scattered information into systems people can reason about.
The technical work matters, but so does what it unlocks: fewer unanswered questions, more confident decisions, and less time spent wondering whether the number is right.
- Bias
- Clarity over cleverness
- Standard
- Reliable by design
- Goal
- Useful, trusted data
Open channel
Have a data problem
worth untangling?
I’m open to data engineering opportunities, advisory roles, and thoughtful collaborations. Tell me what you’re trying to make clearer.