Video by Google DeepMind / Pexels
Urban Infrastructure · Software Engineering · Southeast Asia
Specialist advisory for public agencies, developers, and technology companies navigating complex infrastructure challenges, with software engineering discipline applied to the systems that shape cities.
AI-Augmented Engineering
Claude Code and Gemini CLI are the tools. The architecture, security model, and component decisions are not outsourced — every line generated is read and reviewed before it ships.
The result is software that moves fast without accumulating the hidden debt that unreviewed AI code tends to leave behind.
See the approachClaude Code
Agentic coding in the terminal — full-stack, directed component by component
Gemini CLI
AI-assisted reasoning, code review, and architectural thinking
Secure by Review
Every component read and evaluated — auth, injection, secrets, input validation
CI/CD to Production
Automated pipelines — not prototypes left running, but software that ships
Expertise
Each engagement sits at the crossroads of urban systems and digital technology, serving organisations that need expertise fluent in both.
Strategic guidance on infrastructure systems, from transportation networks to utilities. Technical complexity translated into investment decisions and policy that hold up over time.
Software engineering advisory for large-scale infrastructure projects. From procurement specifications to system architecture, bridging the gap between planners and engineering teams.
Advisory on technology-enabled urban interventions: mobility platforms, civic data infrastructure, and smart city systems, delivered with engineering rigour.
A Distinct Perspective
Most urban planners have never written a line of code. Most software engineers have never shaped a city block. This practice sits at that rare intersection, with hands-on experience building digital systems and working through the realities of infrastructure planning.
The work serves public agencies, developers, and technology companies that need an adviser who can read a masterplan in the morning and review an API specification in the afternoon.
Start a ConversationGeospatial Capability
Geospatial analysis is not merely a supporting tool here. It is a core specialisation. Spatial datasets spanning entire road networks, land registries, grid infrastructure, and urban boundaries are processed automatically and systematically, at the scale of billions of rows.
Powered by Python, Go, SQL, and cloud computing, the geospatial stack is built for production workloads: reproducible pipelines, automated ingestion, and outputs delivered directly to interactive maps or analytical dashboards.
Technology Stack
PostGIS
Spatial SQL queries on relational geometry at scale
DuckDB
In-process analytical queries on massive spatial datasets
Apache Spark
Distributed processing of billion-row geospatial workloads
Mapbox
Interactive mapping and visualisation for end-user delivery
Cloud Infrastructure
Scalable GCP architecture for continuous geospatial data ingestion
CI/CD & DevOps
Automated build, test, and deployment pipelines for geospatial applications
SvelteKit
Frontend framework for building fast, interactive geospatial web applications
Python
Primary language for spatial data processing, analysis pipelines, and automation
Mapping Projects
Interactive maps built and published.
Story Map · Indonesia
An interactive story map narrating the development of Indonesia's first high-speed railway, combining maps, visuals, and editorial narrative.
Accessibility Map · Jakarta
An interactive map visualising rail stations reachable within a given drive time across the Jakarta metropolitan area.
Transit Map · London
A redraw of the London Underground map rendered at true geographic coordinates, revealing the spatial reality behind the iconic schematic diagram.
Bespoke Software
Off-the-shelf tools were not designed with infrastructure planning in mind. Bespoke software solutions are engineered specifically for the workflows, data structures, and decision-making needs of urban planners, infrastructure developers, and public agencies.
From WebGIS platforms and spatial data pipelines to project management tools and analytical dashboards, every solution is built to production standard and deployed on Google Cloud Platform, a specialism held at the professional certification level.
Interactive mapping applications for land appraisal, infrastructure monitoring, and spatial decision-making, built with Mapbox and SvelteKit.
End-to-end pipelines that ingest, transform, and serve spatial and infrastructure data automatically, reducing manual analysis overhead.
Robust backend services in Go and NestJS that power planning tools, integrate third-party data sources, and expose clean APIs for analytical consumption.
Architecture, deployment, and operations on GCP at professional certification level. Scalable, cost-efficient infrastructure for planning applications of any size.
Portfolio
Infrastructure projects contributed to as analyst or adviser.
Software
A WebGIS platform built for solar farm developers to appraise land more effectively. Decision makers can assess local infrastructure, grid connection points, and powerline capacity directly on an interactive map before committing to a site.
Transport
GTFS-based transport modelling for the World Bank to support the feasibility and business case development of Bus Rapid Transit projects in Medan and Bandung, assessing route performance and ridership potential.
Transport
Typology analysis classifying each LRT Kelapa Gading station catchment area by its Transit-Oriented Development character. The classifications inform how the surrounding neighbourhoods should be planned and developed, guiding density, land use, and public realm decisions.
Get in Touch
For those seeking a strategic sounding board, a technology guide, or an embedded adviser, please get in touch to discuss the challenge.
upsutan@gmail.com