AI Engineer

We are seeking an AI Engineer to join a fast-moving technology team building a platform that combines machine intelligence, financial modeling, and large-scale investment datasets to support capital allocation across global markets. The team has already built a suite of core tools that are in internal testing, so this role joins a system with real traction behind it, not a blank slate.

The AI Engineer will build and own the automation that turns firm and client documents, including spreadsheets, PDFs, decks, committee logs, and due-diligence materials, into audit-ready institutional deliverables such as quarterly client reports and investment-committee memos. Every number has to foot, every factual claim has to trace to a source, and a person signs off before anything reaches a client or committee.

This is a high-ownership, small-team environment where the pace is fast, the team interacts multiple times throughout the day, and important releases or business-critical deliverables may occasionally require flexibility outside standard working hours. Explore more opportunities on our IT job openings page., * Build ingestion pipelines across heterogeneous sources, including multi-sheet workbooks, PDFs, and decks with chart data trapped in images, into structured, verified output

• Keep computation, ranking and selection, reconciliation, and compliance checks deterministic, using a model only where judgment is genuinely required and grounding and verifying its output
• Design conservative, traceable entity resolution rather than silent fuzzy matching
• Handle missing, malformed, or irrelevant inputs gracefully, distinguishing “flag and continue” from “stop for human review”
• Build generate, critique, revise, audit, and verify pipelines rather than single-shot prompts, including verifying absence claims against the full source corpus
• Keep domain and analytical knowledge in structured, human-editable data files so the engine stays generic and a different knowledge pack changes the output, not the code
• Design prompt systems as maintainable, composable scopes rather than one large, fragile template
• Contribute to the engines behind investment decisions, including benchmarking, forecasting, and portfolio optimization and underwriting logic
• Work with large-scale, linked investment data at the fund, deal, property, and portfolio level, and keep it connected as new tools consume it
• Design the draft and approve boundary, record how sign-off is captured, and make every material figure and factual claim resolve to a machine-readable source
• Own reproducibility so that the same inputs produce the same output and next period’s inputs work through configuration rather than a rewrite
• Take systems from local scripts to server-deployed production, including environment parity, secrets, scheduling, and monitoring for unattended runs
• Cache expensive third-party extraction calls and re-run only when sources actually change
• Build a lightweight operator interface to kick off runs and review output, with real testing, dependency management, and LLM-call observability
• Translate reporting and analytical rules from investment teams into code and verification logic

• Ability to cleanly separate probabilistic AI behavior from deterministic business logic and validation, and to know what must never be delegated to a model
• Strong engineering fundamentals, with production data pipelines shipped rather than only prototypes
• Experience building LLM-integrated systems with grounding and verification, and a clear sense of where models fail and how to design around it
• Experience with multi-pass and self-critique LLM workflows, and the judgment to know when they are worth it
• Experience extracting data from unstructured and semi-structured documents, including chart data that exists only as pixels
• Demonstrated testing and verification judgment, including how to know output is correct and how to catch a regression
• Comfort taking a system from a laptop to unattended production, including secrets, scheduling, and monitoring
• Clear, honest technical communication about what a system does and does not do, * Asset management, real estate, or other regulated financial services experience
• Vision-capable or multimodal model experience for image and scanned-document extraction
• Third-party document-parsing or OCR API experience and cost-aware caching
• Experience building lightweight internal tools or UIs for non-engineering stakeholders
• Data governance, audit, or compliance exposure
• Information-extraction or taxonomy-induction experience, such as clustering and normalizing inconsistent content
• Familiarity with Azure services such as Azure OpenAI, Azure AI Search, Azure AI Document Intelligence, Azure Blob Storage, Azure Functions or Container Apps, Service Bus, Azure SQL or PostgreSQL, Key Vault, Microsoft Entra ID, and Application Insights (not every service is expected)

Technical Skills for an AI Engineer

• LLM-Integrated System Design with Grounding and Verification
• Multi-Pass Generation and Critique Workflows
• Document Extraction, OCR, and Multimodal Models
• Data Pipeline Engineering and Entity Resolution
• Testing, Verification, and LLM Observability
• Production Deployment, Scheduling, and Monitoring
• Azure (OpenAI, AI Search, Document Intelligence, Functions, Key Vault), * This position is not available for C2C engagements
• Candidates must be authorized to work without current or future visa sponsorship