Software Engineer, Applied AI

Fluidstack is hiring a Software Engineer, Applied AI to build core software primitives and domain models for its AI infrastructure projects. The role also integrates external vendor data and enables AI agents to operate safely through tools, guardrails, and audit trails.

Responsibilities

• Build core primitives end to end. Entity ownership, audit, authorization, orchestration. You make the right thing the default and the wrong thing hard
• Own the domain model. Turn Fluidstack's world of power, datacenters, and chips into composable entities that still hold up years from now
• Define interactions with the outside. Interface with vendor systems and ingest the formats our domain actually speaks: KMZ, BIM, Revit, vendor docs
• Make AI agents first-class operators of our systems. Give them the tools, guardrails, and audit trails to act safely, not just advise

Preferred

• MLE work, especially extraction and synthesis from unstructured data
• Workflow and orchestration engines (Temporal, Cadence, Airflow)
• Geospatial (KMZ/GIS) or BIM (Revit, IFC)
• Authorization and policy engines (OPA, Zanzibar-style, RBAC/ABAC)
• Audit or event-sourcing systems
• Go, TypeScript, and Postgres
• Machine Learning Engineering
• Unstructured Data Extraction
• Unstructured Data Synthesis
• Workflow Orchestration
• Cadence
• Airflow
• Geospatial Data
• KMZ
• GIS
• BIM
• Revit
• IFC
• Authorization Engines
• Policy Engines
• OPA
• Zanzibar-Style Authorization
• RBAC
• ABAC
• Audit Systems
• Event Sourcing
• Go
• TypeScript
• PostgreSQL

Benefits

• Equity, including potentially restricted stock units
• Retirement or pension plan, in line with local norms
• Health, dental, and vision insurance
• Generous PTO policy, in line with local norms

Company

• AI compute infrastructure for labs, governments, and enterprises
• New York, New York, USA
• https://www.fluidstack.io/
• Founded in 2017
• 501-1000 employees

Qualifications

• You move toward ambiguity, not away from it. You walk into the fog, build the map, and explain it to everyone else
• You learn at a steep slope. You reach real competence in an unfamiliar domain fast, whether it's a workflow engine, an authz model, or a geospatial format. We value this over existing expertise
• You have taste in abstraction. You see the one primitive hiding under five janky one-offs, and you know when not to abstract
• You model domains with rigor earned the hard way. You've gotten it wrong, felt the pain, and never forgot it
• You pull structure out of mess. You've done real work turning messy, unstructured data into something usable, in ML, data engineering, or research
• You're fluent with AI tooling. LLM APIs, MCP servers, and agentic frameworks, and you drive Claude Code, Cursor, or similar every day
• You've shipped production code that other people depend on, and you're comfortable in any language using AI coding tools
• MLE work, especially extraction and synthesis from unstructured data
• Workflow and orchestration engines (Temporal, Cadence, Airflow)
• Geospatial (KMZ/GIS) or BIM (Revit, IFC)
• Authorization and policy engines (OPA, Zanzibar-style, RBAC/ABAC)
• Audit or event-sourcing systems
• Go, TypeScript, and Postgres

Required skills

• Agentic Frameworks, Airflow, Cadence, Claude Code, Cursor, Domain Modeling, Event Sourcing, Go, IFC, KMZ/GIS, LLM APIs, MCP Servers, OPA, Postgres, Production Software Development, RBAC/ABAC, Revit, TypeScript, Unstructured Data Processing

Preferred skills

• Audit or event-sourcing systems, Geospatial (KMZ/GIS) or BIM (Revit, IFC), Go, TypeScript, and Postgres