Data engineering & analytics · Cairo, Egypt · with Denver, US
Full name: Mohammed Nabil
I build data platforms, analytics and AI that whole companies rely on, and hired and lead the engineers behind them. Two recent wins: an AI invoice audit and economic forecast comparisons (reserves), each cutting reviews from days to hours. Open to Data Engineering Manager, Head of Data, and Staff/Senior AI Data or Analytics Engineer roles.
Moves with the bit · click a stop to jump there
At a glance · 2004 – today · Cairo, working with Denver
Long experience in software development, data platforms, analytics engineering and machine learning. Each stage added to the one before it: software and spatial solutions first, then the data platform and the team behind it at Raisa Energy, and today AI that answers business questions from the warehouse.
The log beside this shows the whole career at once, one layer per stage.
Scroll to go through it step by step ↓
Cairo University · 2004
It starts at Cairo University1, where I finished a Bachelor of Computer Science in 2004.
Everything after builds on that foundation: software first, then spatial solutions, then data platforms and AI.
Technical Lead / Senior Developer · Esri Northeast Africa, Cairo
At Esri I built GIS systems, software whose first question is “where is it?”, for clients in several countries. Among them, vehicle tracking for a municipal fleet2 and a spatial homeland security system3.
As technical lead I directed a group of engineers4 across front end, back end and spatial data analysis, and ran the Agile Scrum process.
GIS Development Manager · Link Development, Cairo
I owned the technical direction of GIS development, mentored one direct report5 and coordinated with engineers in other teams.
The work widened from GIS systems to spatial data analysis combined with machine learning6, with location as one more signal in the models.
Staff Data Scientist (Data Engineering & ML) · Raisa Energy, Cairo / Denver
At Raisa Energy I kept spatial analytics in my toolkit and built on it. I ran a hands-on evaluation of Snowflake against Databricks, chose Snowflake, and introduced Snowflake and dbt7 as the company's data platform, which I then designed end to end.
I built ETL / ELT pipelines and cloud data infrastructure for the data science team, a framework to deploy models with a simple script8, and classical machine learning models9. I also generated synthetic well paths for wells with missing survey points, and turned permitted well locations from text into a spatial layer.
Data Engineering Manager · Raisa Energy
I was promoted to build Raisa's first dedicated data engineering team10, separated from the combined data science and engineering team. I hired and onboarded the engineers and led them to build the data warehouse on Snowflake11.
Director of Data Engineering · Raisa Energy, Cairo / Denver
I lead a team of up to seven engineers12, a team I built from scratch, and still ship hands-on in architecture, design and code. I'm responsible for hiring, compensation and the team budget, and contribute to company-level budget planning.
I shape strategy, vision, roadmap and OKRs / KPIs13 with the company leadership, work remotely with the reserves, subsurface and accounting teams in Denver, and introduced regular R&D weeks14 where the team evaluates new technologies.
The Snowflake warehouse
I designed the warehouse in layers15: a custom medallion architecture with dimensional modeling, used by several teams across the company. Abstraction layers make complex sources easy to query, combining well headers and production volumes from several data providers16 into a public well data layer, and joint interest bills with revenue, costs and accounting into a financial layer, with data marts for each team on top. Spatial pipelines in PostGIS and ArcGIS keep well locations next to the well, production and financial data.
Quality is built in: role-based access in Snowflake and dbt tests17, documentation and lineage, monitoring, alerts, query tuning and cost management.
AI in production
I built an MCP server and semantic layer on Snowflake Cortex18 so AI assistants can answer business questions from warehouse data. The company depends on it now.
An AI-assisted audit of joint interest bills19 and reserves run snapshots20 each cut a review from days to hours. I also built AI agents and LLM tools21 for reporting, text-to-SQL and document extraction, and mentored an intern who built an agentic reporting tool.
The work continues
Below: key deliverables, skills and education.
Director of Data Engineering · Raisa Energy · 2023 – present
What the platform delivers today.
Built the team from scratch to up to seven engineers and still ship hands-on. Set standards for naming, modeling and documentation, review code and coach by asking questions, and run R&D weeks and knowledge-sharing sessions.
Result: Own hiring, compensation and the team budget
A layered Snowflake warehouse (staging, intermediate, marts, BI) with dimensional modeling. Shared core layers unify public well data (headers, production volumes) and join joint interest bills with accounting, and each team gets its own data mart.
Result: About ten source systems · used by several teams
A config-driven framework loads only changed tables and keeps full history. dbt tests act as gates between layers, with role-based access, lineage, monitoring and cost tuning in Snowflake.
Result: Users keep the last good data when a test fails
Scoring models and indicators that turn engineering, financial and production data into one number. Example: the bounding score, built from nearby wells, their distance, cost and production, shows how surrounded each well is and how much production is at risk when new wells are drilled close by. It is used to approve or reject funding for new wells. Model runs are snapshotted and compared on one dashboard.
Result: Model-run comparisons days → hours
An MCP server and semantic layer on Snowflake Cortex with Azure AD sign-on and role-based access. AI agents handle reporting, text-to-SQL and document extraction, and an AI-assisted audit runs about 40 checks on every joint interest bill (JIB) before payment, with fixed checks in Snowflake and AI only where judgment is needed.
Result: Company-wide daily data questions · invoice review days → hours
Skills
Every tool I've put to work, grouped the way the CV groups them. The last group lists tools I've evaluated hands-on.
Learning
Computer science first, then management of technology.
Nile University, Graduate School of Management of Technology, Cairo
Cairo University, Faculty of Computers and Artificial Intelligence, Cairo
What comes next
What I write, and the small projects I build and host. New ones appear here.
Contact
For questions about data platforms, analytics or anything on this page, reach me by email or on LinkedIn.