AI Operations Engineer
Coast · New York, NY
About this role
**AI Operations Engineer — Coast** ## About Coast Coast is re-imagining the U.S. B2B card payments infrastructure for the country’s 500,000 commercial fleets, 40 million commercial vehicles, and millions of commercial drivers. Coast builds modern payments and spend management for fleet businesses—fuel cards, corporate cards, and bill pay—while becoming **AI-native**, so repeatable work is handled by AI and teams can focus on customers and hard problems. You’ll join a small internal AI team as one of **two AI Operations Engineers**, reporting to the **AI Lead**. ## About the Role This is a builder role with an operator’s mindset. You’ll find high-leverage work across **Sales, Customer Success, Marketing, Support, Finance, Risk, Product, and Operations**, and own it end to end—from first conversation to production. Some weeks you’ll write code; others you’ll run workshops or redesign processes with teams. Everyone at Coast uses Claude and builds their own skills. You’ll build what others can’t: **autonomous agents** and the foundation that improves everyone’s workflows. **First 6 months:** ship **3+ agents** that teams rely on daily, each backed by **evals** showing month-over-month improvement, and drive measurable operating efficiency gains (tracked in **hours saved**). ## What You’ll Do - **Find and prioritize the work:** Sit with teams, map workflows, estimate time saved, and build a ranked backlog with the AI Lead. - **Build autonomous agents:** Ship event-driven agents that perform multi-step work across the stack. You’ll own: - **Loop & orchestration:** agent loops, sub-agent handoffs, and when to act alone vs. escalate to a human. - **Memory & context:** shared memory as a knowledge graph; decide what each agent sees, remembers, and forgets. - **Custom tooling:** build the tools agents need. - **Traces & evals:** trace every run, score outputs against test sets from real outcomes, and prove each version beats the last. - **Measure impact:** Define success metrics, track adoption and time saved, and report results to leadership. - **Run what you ship:** Own reliability, latency, and cost in production; catch regressions early. Partner with Security and IT on access, audit trails, and sensitive data handling. - **Enable the company:** Build shared infrastructure (knowledge bases, application databases, templates, tools, connectors). Support adoption via training, office hours, hackathons, and an AI champions network. - **Stay on the cutting edge:** Evaluate new models/tools as they land and raise the bar on what you build. ## Projects you might take on - **Autonomous product help docs:** Capture questions from customers (support tickets/calls) and employees (Slack), detect missing/outdated articles, draft updates, route for review, and learn from what gets approved and shared. - **Deal coach for large trials:** Build an agent that creates value calculators and tailored pitch decks from call transcripts and account data, coaches AEs/CSMs through trials with next steps, and learns which plays win. ## Requirements - **2–5 years** in operations, product, solutions, forward-deployed engineering, internal tools, or a similar role at a fast-moving company. - Shipped **AI workflows/automations** that people use every day—show what you built. - Hands-on fluency with **Claude Code, Cursor, or similar coding agents**, plus LLM tooling (prompt design, tool use, **MCP**). - Built at least one agent with **persistent memory** and/
Listing freshness
CronJobs last confirmed this listing 1h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.