Role overview
CDS's 2026 strategy explicitly calls out AI as a core lever across multiple departments:
- PH1/PH2: Implement AI to analyze municipalities, review reports, flag areas of concern, and organize permitting information.
- Estimating: Use AI to improve accuracy, minimize scope gaps, and eliminate errors in bids
- HR / Recruiting: Build an AI workflow to screen and qualify candidates at scale using a scoring system
- Finance: Automate vendor/supplier AP tracking and cash flow modeling
- Business Development: Support outreach workflows, portfolio updates, and client onboarding processes
You will be the person who builds, deploys, and maintains all of it.
*KEY RESPONSIBILITIES
Agent Design & Deployment**
- Design, build, and deploy AI agents tailored to specific departmental workflows across CDS
- Develop and maintain custom GPTs, Claude projects, and similar LLM-based tools using platforms such as OpenAI, Anthropic, and Cohere
- Architect multi-step automation workflows using Make.com, Zapier, n8n, or similar orchestration tools
- Build intake forms, triggers, and conditional logic to route information to the right agent at the right time
What we're looking for
| AI Platforms | OpenAI (GPT-4, custom GPTs, Assistants API), Anthropic Claude, Gemini; ability to build, configure, and deploy agents on each
| Prompt Engineering | Expert-level prompt design for instruction-following, structured output, classification, extraction, summarization, and multi-step reasoning tasks
| Automation Tools | Make.com (Integromat), Zapier, n8n — building multi-step workflows with branching logic, error handling, and data transformation
| APIs & Webhooks | REST API integration, JSON/XML parsing, OAuth, webhook configuration; ability to connect AI agents to third-party platforms without native connectors
| Programming | Python (required): scripting, data processing, API calls, file manipulation. JavaScript a strong plus. Ability to write clean, functional scripts — not production engineering, but reliable automation code.
| Data Handling | Spreadsheet logic (Excel/Google Sheets), basic SQL or structured data querying, JSON/CSV manipulation, data cleaning and transformation
| AI Agent Frameworks | Familiarity with LangChain, CrewAI, AutoGen, or similar multi-agent orchestration frameworks; understanding of RAG (Retrieval-Augmented Generation) architectures
| Tool Integration | Google Workspace (Gmail, Drive, Docs, Sheets), Slack, Procore, Smartsheet, Harvest, or similar PM/construction platforms
| Vector Databases | Basic understanding of embeddings, vector search (Pinecone, Weaviate, Chroma), and knowledge base retrieval for document-aware AI agents
| Problem Solving | Ability to sit with a department head for 30 minutes, map their workflow, identify where AI creates leverage, and propose a build plan the same day