OPERATIONAL
Six+ years of GIS expertise.
Currently building at the seam of maps and intelligence.
A short manifesto · Vol. 01

Every map tells a story.
The next ones will think.

I'm Amar Chandrasekhar — a GIS professional working at the seam where geospatial data meets intelligent software. Six+ years of GIS expertise inside Esri, ArcGIS, and the geodatabases that quietly run cities, utilities, real estate, and government. Now building toward the layer everyone else is ignoring.

Read the thesis Get in touch
6+
Years of
GIS expertise
3
Sectors
shipped in
2
Continents
of fieldwork
1
Wedge worth
betting on

For most of recorded history, we surveyed the world with chains, drew it on paper, and trusted the people who walked the lines.

Then came GIS. Then came Esri. Then came the geodatabase — and for the first time, the map became a living record instead of a static one.

The next era won't be defined by who has the most data. It will be defined by whose maps are intelligent, whose assets are self-documenting, and whose decisions happen in software — long before they ever happen in the field.

And yet — most GIS today is still treated like a filing cabinet. A static record of where things sit. A place where outages are logged, parcels are stored, and dashboards quietly go to die. Esri gave us the cathedral. We've spent two decades using it as a closet.

That's the gap. The geospatial system of record is the most underrated layer of the modern stack — the place where every asset, every outage, every parcel, every permit, every decision converges — and almost nobody is building intelligence into it.

I am.

Amar Chandrasekhar · GIS · ArcGIS · the geodatabase & what's next
// What I'm building

The work that doesn't fit on a résumé. Tools, prototypes, and small bets — all aimed at the same idea: making spatial data think. What's leading the table right now —

In Build · Solo PROJECT // ATLAS-01

An AI co-pilot for the geodatabase.

A small, opinionated tool that sits next to ArcGIS Pro and reads your enterprise GDB the way a senior GIS analyst would — surfacing schema drift, orphaned features, broken connectivity, and inconsistent domains before they become field tickets, permit conflicts, or outage misroutes. Built on the QA/QC patterns I've run by hand for years across utility, real estate, and government datasets — now wrapped in a language model that understands what "wrong" looks like in real Esri data.

The wedge: every organization with a serious GDB runs the same broken cleanup loop. Analysts spend half their week clicking through feature classes. That work is going away. The question is who builds what replaces it.

stack — Python · ArcPy · Esri REST · LLMs · ArcGIS Enterprise · Postgres + SDE
Concept · In design

A natural-language layer for ArcGIS Online.

"Show me every transformer installed before 2010 within 50 feet of a flood zone." Today that's a model. Tomorrow it's a sentence. The geodatabase deserves a real interface.

Concept · In design

Self-documenting feature services.

Every layer published from Portal should ship with its own provenance, lineage, and quality score baked in — not as a separate document nobody reads, but as live metadata the field team can trust.

Open invitation

Your idea, maybe.

Building something at the seam of GIS, AI, and spatial infrastructure? Let's talk. The interesting work is happening at edges most people aren't watching.

// Track record

Things I've already shipped that prove I can do the harder thing next. Six+ years, three sectors, the same obsession: making spatial data tell the truth.

i.
Migrated a legacy electric distribution network into a fully-modeled ArcGIS Utility Network — connectivity, topology, domains, the whole stack.
Energy client · Texas · ArcGIS Pro · Utility Network · Python · SQL · OMS / SCADA
+35% / −40%
RELIABILITY · LOOKUP TIME
ii.
Built a spatial scoring engine for a real estate developer — turning raw parcel data into ranked, development-ready land in days instead of weeks.
Real estate developer · Texas · ArcGIS Pro · QGIS · Survey123 · Multi-criteria Analysis
−50%
SITE SELECTION TIME
iii.
Currently shipping field-grade web applications for utility crews — Experience Builder apps that put the right asset data in the right hand at the right moment.
Energy client · Texas · Experience Builder · ArcGIS Online · Enterprise GDB · ArcPy
now.
iv.
Earlier — cadastral surveying for a state government across multiple districts. The kind of fieldwork that teaches you what data really is.
Government of Andhra Pradesh · India · DGPS · Total Station · ArcMap · Cadastral GIS
4yr
FOUNDATION
// What I reach for

The Esri stack is home — ArcGIS Pro, Enterprise, Online, Portal, Experience Builder, Field Maps, Survey123, StoryMaps. I've shipped real things in all of them.

For the data layer I live in ArcGIS Utility Network, Geometric Network, SDE multi-versioned editing, and enterprise geodatabase design. I've broken and rebuilt enough GDBs to know which schemas survive contact with reality.

For automation and intelligence: Python, ArcPy, SQL, Esri REST, and increasingly LLMs pointed at spatial data. The boring parts of the job are about to get a lot less boring.

Field-side I came up on DGPS, GNSS, and total stations. Most of what I believe about clean data, I learned standing on a transect with a survey rod. That perspective doesn't go away.

// Field notes

Four things six+ years inside GIS have made me believe.

i.

The geodatabase is the most underrated piece of software in any organization that touches the physical world.

Every utility outage trace, every permit decision, every site selection, every cadastral dispute, every parcel valuation — all of them quietly depend on a feature class somewhere being correct. Get it right and the rest of the stack flies. Get it wrong and no amount of dashboards will save you. The map isn't a deliverable. It's the operating system.

ii.

Esri built the cathedral. Most teams are using it as a closet.

ArcGIS Enterprise can do things most analysts have never asked of it — real-time services, network tracing at scale, federated portals, REST endpoints feeding entire applications. The platform isn't the limitation. The imagination layer on top of it is.

iii.

AI won't replace the GIS analyst. It'll free the GIS analyst to finally be a designer.

Sixty percent of the job is schema mapping, attribute validation, and writing the same QA scripts in slightly different forms. That work is automatable now. What remains — knowing what "correct" looks like in a particular dataset, designing the systems, choosing the wedge — is human, strategic, and about to become more valuable, not less.

iv.

The next great organizations — in any sector — will be the ones that treat their geospatial data as infrastructure, not overhead.

Watch which utilities, which cities, which developers, which agencies invest in their geospatial system of record over the next five years. Watch which ones treat their GDB as a strategic asset. That's the leaderboard nobody else is reading yet.

// Working with me

Let's make spatial data think.

I take a small number of conversations seriously. The ones I'm most interested in: founders building at the seam of GIS, AI, and spatial infrastructure. Organizations that want to treat their geodatabase as a strategic asset instead of a filing cabinet. Anyone with a sharp problem in spatial data who wants a builder, not a vendor. The door's open. Pick a thread.