Check the address, radius and neighbourhood boundaries. They describe different areas.
Making Public
Data Practical.
I design and build products that make complex public data useful. I founded Perimtr to turn fragmented municipal records into clear, address-specific reports.

Perimtr.
Evidence, not verdicts.
Public data is fragmented across portals, registries, and APIs. Perimtr binds these signals directly to a specific address—making sources, spatial coverage, and data limits immediately actionable.
A neighbourhood report grounded in public records.
See how a single address aggregates hyper-local geographic context, active permits, 311 service logs, and municipal risk factors in one clear view.
How to read
a finding.
Before a number becomes a conclusion, ask four questions.
349 active permits within 1,000 metres means matching records—not 349 construction sites.
Look for the source’s reporting period and update date before comparing figures.
Read the named source, method and limitations. Missing records do not mean nothing happened.
Many sources.
One accountable system.
From fragmented civic records to structured address reports, every step is explicit, traceable, and reviewable.
Acquire
Ingest civic datasets from open-data portals and public APIs.
Validate
Audit schemas, track source revisions, and verify geographic and time coverage.
Locate
Resolve the address, map proximity and 1,000-metre areas, and measure overlapping neighbourhood boundaries.
Interpret
Connect relevant records while preserving source provenance and uncertainty.
Report
Generate readable, source-traceable reports around a specific address.
Test and operate
Run regression checks, review outputs, release verified builds, and monitor the live service.
Same label.
Different places.
An address search can return a result while pointing to the wrong location. I audited 3,165 repeated address-label groups in Perimtr’s 525,401-point Toronto index. Select a distance to see what the investigation revealed.
Maximum separation between points sharing an address label
520 groups contain points no more than 5 metres apart. Close locations still need careful matching.
AI is the engine.
I am the architect.
Perimtr is built on non-negotiable product decisions—defining what questions matter to real estate, what data constitutes proof, and how to rigorously handle uncertainty.
Spec → Prompt → Review → Verify → Ship
I write exact rules, orchestrate multi-agent execution, validate output against raw public records, investigate edge-case failures, and gate every deployment.
684 collected.
681 passed. 3 skipped.
The tests check behaviours across data, address resolution and report logic. For example, this run checks that negative walking distances raise an error and unknown neighbourhoods are handled explicitly.


Aerospace precision.
Artistic clarity.
My aerospace background taught me to examine how complex systems behave under pressure. Advertising art direction taught me to make information clear to the person using it.
Perimtr brings those disciplines together. I scrutinize the data pipeline, test where it can fail, and shape the report so people can understand the evidence and its limits.
B.Tech, Aerospace specialization · Postgraduate Certificate, Advertising Art Direction, Humber Polytechnic
Let's talk about
what comes next.
Interested in how I approach complex product and data problems? I’d be glad to show you the product and discuss the decisions behind it.