AI/ML Engineer

About the role

ABOUT FUTURE WORKS

Large companies spend billions on consulting and implementation projects to adopt new technology. Most of that work is slow and expensive, and too often it fails to deliver. Future Works replaces it. We rebuild the operations that move a company's P&L in twelve-week cycles, run by senior experts and fleets of AI agents, with our fees tied to outcomes our clients' own finance teams verify. We deliver in twelve weeks what incumbents quote at two years.

We work with large companies in energy, construction, real estate and other asset-heavy industries. Every role here is built around humans and agents working as one system. Whatever the seat, we look for high energy, high agency and a real interest in building with AI.

THE ROLE

We are running an engagement to design and build an AI agent system that manages a complex, high-value transaction workflow for a large enterprise client, from initial engagement through close-out, across a multi-stage process.

As an AI/ML Engineer, you will build agent capabilities in parallel during the build phases, working under the Lead AI/Agent Architect. You will implement the tool integrations, data pipelines, and per-stage agent components that let a full transaction scenario run end to end against sandbox data.

KEY RESPONSIBILITIES

- Per-Stage Agent Build: Implement agent capabilities for individual workflow stages (for example engagement agreement generation, needs-assessment brief, market survey and long-list scoring, tour management, RFP and proposal analytics, lease abstraction, and close-out and invoicing), building to the orchestration model and stage handoff protocols defined by the Lead Architect.

- Tool & Integration Development: Build the typed tools the agents call: template population, market-data query, financial analytics (NER, TI, free rent, NPV), clause and lease-clause extraction, and document assembly.

- Data Pipelines: Build and maintain pipelines against sandbox and representative sample data, including the property availability data source.

- Human-in-the-Loop Implementation: Encode the approval gates and review checkpoints specified by the architect, so binding steps always route through a broker confirmation.

- Quality & End-to-End Testing: Validate components against deal-scenario data and participate in end-to-end testing across all eight stages, checking artifact quality and stage handoff integrity.

- Build-Phase Readiness: Participate in Phase 1 architecture workshops to ensure the build starts on solid foundations.

WHAT WE'RE LOOKING FOR

REQUIRED SKILLS AND EXPERIENCE

- 4+ years in software or machine learning engineering, with production experience building LLM-powered features or agents.

- Agent Tooling: Hands-on experience with agent frameworks, tool and function calling, retrieval, and prompt engineering.

- Python & Data Engineering: Strong Python and practical experience building data pipelines and integrating structured and unstructured data sources.