We are seeking a Research Software Engineer (GeoBIM and 4D Systems) focused on researching core technologies that enable versioned zone governance, a spatiotemporal data hub, LLM-ready context packaging (MCP), and enterprise ArcGIS interoperability.
Responsibilities
Turn problem statements into schemas, interfaces and reference implementations for versioned zones, 4D spatiotemporal linkage, MCP read-only tools, schedule normalization and ArcGIS exchange.
Define governance rules for user-defined zones (geometries) including merge, lock and approval. Model lineage and reporting semantics that remain stable as zones and plans evolve.
Establish controlled vocabularies, validation checks and reconciliation outputs for mapping schedule activities to zones and BIM components.
Implement retrieval endpoints and fact blocks for LLM consumption with clear provenance and redaction rules.
Prototype publishing and ingest workflows to and from enterprise GIS. Write runbooks and acceptance checks.
Technical Expertise
Data and APIs: Python, REST services, data modeling, Postgres or PostGIS, and caching with tools like Redis
Geo and BIM: BIM concepts, 3D assets such as point clouds and meshes, common geo formats and projections, and enterprise GIS patterns
4D data hub and MCP: schema and ontology design for provenance, freshness and confidence, plus read-only MCP tools for retrieval
ArcGIS interoperability: FileGDB, GeoJSON, shapefile and REST Feature Services publishing and ingestion
Schedule normalization: P6 to Excel mapping, controlled vocabularies, validation checks and reconciliation reports
Security and provenance: auditability, reproducibility and evidence-backed analytics in enterprise environments
(Optional): internal admin consoles with TypeScript and React and LLM or GenAI frameworks such as LangChain, LlamaIndex or Haystack
Must-Have Qualifications
Proven delivery of end-to-end progress tracking or similar pipelines from ingestion to write-back
Experience implementing multi-image instance association using camera poses and geometric checks
Ability to design majority-voting status logic with evidence weighting and monotonic completion
Practice building occlusion and coverage scoring that generates actionable capture-gap outputs
Capability to specify and prototype BIM delta engines with lifecycle states and cross-version ID reconciliation
Strong Python engineering discipline, testable APIs and reliable data contracts
Good-to-Have Qualifications
Hands-on operational work with Revit exports and parameter handling and with Autodesk Construction Cloud projects and permissions
Proficiency with IFC and FBX interchange and basic familiarity with GLB or OBJ
Experience with VLM or RAG style retrieval and metadata policy
Familiarity with ArcGIS Enterprise publishing patterns and schema design
Construction and BIM exposure across trades and progress reporting
Light front-end experience for internal tools using TypeScript and React