01
PostgreSQL & PostGIS
Database design and performance tuning for global spatial datasets: partitioning, spatial indexes, query plans and the maintenance that keeps production systems useful.
Database engineering at GeolytixGeospatial data engineering / Leeds, UK
I build geospatial data systems that work in production.
I'm Omar, a data engineer based in Leeds, with more than a decade across geospatial engineering and commercial data products. I build PostgreSQL/PostGIS databases, spatial pipelines and cloud services for retail location planning, 3D city data and demographic analysis.
A bit about me
I like the messy bit between data and a useful thing.
My work connects spatial data engineering with commercial delivery: global location data at Geolytix, 3D data services at VU.CITY and demographic products at CACI.
I work across database design, spatial processing, APIs and cloud infrastructure, and with product and commercial teams to scope datasets and coordinate delivery. I like the engineering detail, but I also care about what the system helps people do.
Expertise
From spatial data to a working product.
02
Spatial models & analysis
H3, SQL and Python workflows for catchments, distance calculations, spatial joins and hex-based analysis behind retail and location-planning products.
Spatial processing at Geolytix03
Data pipelines & cloud services
FME and Python pipelines from ingestion through validation and publishing, backed by cloud PostGIS, APIs and services on GKE. Reproducible jobs and CI/CD carry the work into production.
Cloud data delivery at VU.CITY04
AI, with guardrails
Exploring AI-assisted mapping in Portico through structured tools, semantic metadata, scoped permissions and verification. An independent platform currently in development.
Explore the Portico projectProfessional experience
Geospatial engineering with commercial purpose.
Data Engineer / Geolytix
Designing and maintaining production geospatial databases for global retail and location-planning products.
- Design and tune PostgreSQL/PostGIS schemas, partitioning and indexes for global spatial workloads, with query-plan analysis and database maintenance.
- Build H3, PL/pgSQL and Python workflows for catchments, distance calculations and spatial joins used in location analysis.
- Turn data-science code into customer-facing production systems, optimising SQL and Python and building reproducible data flows with CI/CD for scheduled jobs and controlled releases.
GIS Data Scientist / VU.CITY
Built the automation, APIs and cloud data services behind 3D spatial data delivery and model ordering.
- Built FME and Python ETL workflows covering ingestion, validation and publishing of 3D datasets, alongside internal web services and APIs.
- Managed cloud PostgreSQL/PostGIS systems, supported services on GKE and developed pipelines for Unity asset generation.
- Delivered a map-based 3D model ordering system, integrating internal APIs and spatial data services to simplify requests, drive noticeable sales growth and halve the time from order to deliverable.
Senior Consultant / CACI
Built and managed demographic data products across EMEA, connecting data processing with product and commercial delivery.
- Created high-volume FME workflows and reproducible pipelines for regular demographic data refreshes.
- Worked with product and commercial teams to scope datasets, communicate roadmaps and coordinate delivery.
Selected work
The problem, the engineering, the result.
Geolytix / Production data systems
Geospatial databases & spatial processing
- Problem
- Retail and location-planning products depend on global spatial datasets and repeatable location analysis.
- My contribution
- Design and maintain PostgreSQL/PostGIS databases, tune partitioning and indexes, and build H3, PL/pgSQL and Python workflows for catchments, distances and spatial joins.
- What it supports
- Customer-facing location-planning systems, with data-science workflows brought into production through reproducible data flows and scheduled jobs.
VU.CITY / Commercial spatial product
Model Ordering System
- Problem
- Customers needed a simpler way to specify the areas they wanted and request the corresponding 3D models.
- My contribution
- Built a map-based area selection and ordering system, integrating internal APIs and spatial data services with the model request process.
- Result
- Drove noticeable growth in model sales and halved the time from order to deliverable, while simplifying requests and reducing back-and-forth.
A couple of personal builds, too.
Personal build
Salary, Bonus & Pension Optimiser
A UK compensation modelling tool for comparing cash-versus-pension bonus trade-offs, built as a responsive interactive experience.
Open projectJust for fun
Bead Blast
A colourful little abacus game for a five-year-old: move the beads, hit the target, celebrate the wins.
Play the gameHave a look through my repositories.
Independent work / In development
Portico: exploring AI-assisted mapping.
Portico
Portico is an experimental mapping platform I am building around both human and AI use.
An AI can inspect data and configuration, find out what tools are available, suggest a change and check the result. It does not get unrestricted access to the database or cloud account.
The rule of thumb
Give it enough context to help. Keep the risky bits visible and reversible.
structured toolssemantic metadatareview before changeverify after change
Got a knotty data or mapping problem?
I am open to conversations about data engineering, geospatial systems, cloud infrastructure and sensible ways to use AI with the systems that already matter.