Forward Deployed Engineer

About the role

ABOUT US

Higharc is a VC-backed startup that is changing how new homes are designed and built. Join a founding team who’ve shipped products for Autodesk, Electronic Arts, Nike, and Apple. We have raised over $175M with support from top-notch venture capital firms and more than 18 strategic investors: industry leaders in construction, building products manufacturing, and distribution.

Higharc is seeking a Forward Deployed Engineer to join as one of the founding roles on our new Forward Deployed team. You'll work embedded alongside a large production homebuilder to turn plans into actionable data.

WHAT YOU'LL DO

A homebuilder's plan library is the product data behind its entire business — and at enterprise scale, that data is almost entirely unstructured. Plans live as drawing sets scattered across semi-autonomous divisions, renamed and re-numbered as they move. Nobody can say with confidence which plans are duplicates, how many versions of a given home exist, or what actually varies between them.

Higharc converts those catalogs into structured, queryable plan data, and builds the analysis on top of it: searchable plan libraries with agent-enabled natural-language search, geometric similarity scoring that surfaces plans which have forked under different names, structured mapping of what varies across a library, and classification that makes consolidation decisions possible. You'll build what gets shipped in the field.

This role is for someone who wants to build against real, messy data in front of the customer, not two abstractions removed from it.

Expect to:

- Embed with HQ, regional, and division teams to run technical discovery — profile plan assets, assess data quality, and reverse-engineer how each division's drawing sets encode variation

- Build and ship the searchable plan library, analytics dashboard, and agent-enabled natural-language search — own the pipeline from raw plan ingestion through structured, queryable output

- Tune and extend Higharc's Autotranslate AI capabilities to identify, categorize, and translate floorplans into data

- Work hands-on with the data — write the ETL, build against internal APIs, and ship working capability back to users directly rather than handing off a spec

- Build geometric similarity scoring and classification models that surface plans which have forked under different names

- Meet delivery milestones — coverage targets, data-field completeness, QA error thresholds — and flag tradeoffs in real time when schedule pressure argues against them

- Instrument and quantify results: redundancy across a catalog, dimensional spread on repeated conditions, consolidation candidates, and the purchasing leverage each one represents

- Partner with the Forward Deployed Product Manager, Technical Lead, QA, Intake Specialists, and Client Strategy, and present technical progress and tradeoffs to executive stakeholders

- Build the repeatable version — turn one-off scripts and data patterns into reusable tooling that lets the forward deployed model work on the next engagement, not just the current one