Solutions Architect (Post-Sales, Agent Deployment)

New York Or Bostonfull-timemid

EngineeringSoftware Architect

0 views0 saves0 applied

Apply Now

Quick Summary

Key Responsibilities

triggers, tools, approval gates, override paths, dry-run modes Decompose real-world finance workflows into agent skills,

Requirements Summary

prompt engineering, agent architecture, and workflow design. Demonstrated experience deploying agents or LLM-powered workflows for real users — production, not prototypes.

Technical Tools

EngineeringSoftware Architect

Adaptive is a vertical AI company of 50, building the agentic operating system for the $2 trillion U.S. construction industry. We work with over 1,000 customers, partner with over 50 accounting firms, and have raised $26m from Emergence, a16z, and other top funds.

What We Offer

~1 min read

Win a massive category that's up for grabs. Construction is among the least digitized industries left, and no AI king has been crowned. With our momentum, team, and backing, we are in position to be that company

Build a moat that compounds. We orchestrate multi-stakeholder workflows across an industry that today runs on paper, email, and phone calls. Every process we manage generates data no one else has. That data makes each new use case easier to expand to, while building a compounding moat

Work with the rare team that can pull this off. Winning here takes deep construction knowledge and deep AI fluency in the same room. That intersection is rare, and we have it with backgrounds across finance, technology, and construction

Serve the people building the world around you. Your favorite local café. The new house on the corner. The massive new skyscraper at Central Park South. All built by our customers: skilled, creative people who have spent their lives bringing neighborhoods to life

✓Competitive salary plus meaningful equity

✓Comprehensive health, dental, and vision insurance

✓401(k) match and flexible PTO policy

✓A seat at the table in a rapidly scaling vertical AI startup

✓Unlimited opportunities for professional growth and leadership

About the Role

~1 min read

This is a forward-deployed role. Once a customer signs, you are the person who makes the AI real: configuring, testing, and launching the agents that automate their finance operation. Compliance checks, payment gates, COI tracking, AR collections, billing automation, WIP and forecasting workflows. You sit inside the customer’s actual operation, decompose their messy human processes into agent logic, and own the deployment through go-live and beyond.

Responsibilities

~1 min read

→
Design and configure AI agents for production customer environments: triggers, tools, approval gates, override paths, dry-run modes
→
Decompose real-world finance workflows into agent skills, and make the judgment calls on where automation is reliable versus where it needs a human in the loop
→
Test, debug, and iterate: when an agent produces a wrong output, trace whether the failure is in the prompt, the data, or the process design
→
Build integrations and data pipelines between customer ERPs and Adaptive, including one-off migrations from legacy systems
→
Validate financial output: trace agent-produced numbers back to source and fix the logic when it is wrong
→
Run onboarding through go-live: prerequisites, configuration, training, and the first month of production operation
→
Document what you build so every deployment makes the next one faster

Hard requirement — AI fluency and agent design. You build with AI, not just talk about it. Deep hands-on experience with LLMs as a working instrument: prompt engineering, agent architecture, and workflow design. Demonstrated experience deploying agents or LLM-powered workflows for real users — production, not prototypes. We will test this in the interview.

Structured thinking about data: expert Excel, comfort reading API docs and data models, SQL or scripting a plus
Construction finance literacy (WIP, percentage-of-completion, job costing, AIA billing, retainage, COI compliance) or the domain depth to learn it fast and hold your own with a controller
Familiarity with construction ERPs (Sage Intacct, Sage 300 CRE, Vista, Spectrum, Acumatica) and Procore’s data model is a strong plus
Customer-facing composure: you can run a working session with a controller, manage expectations through a rocky deployment, and communicate clearly in writing
High ownership and comfort with ambiguity. You define the playbook rather than follow one