AI Solutions Engineer (Customer-Facing) - AI Agents, Prompts, Evals & Client Enablement

TO APPLY: Record a 1-minute video using Loom.com introducing yourself and your relevant experience, and include a link to or a short demo of one AI agent, assistant, or automation you have built yourself. LeadCenter.AI is an AI-powered CRM built for financial advisors. Our clients use Troy AI, our agentic AI workforce, and Claudia, our AI assistant, every day to work their leads, clients, and pipeline. We are hiring a hands-on, customer-facing AI Solutions Engineer to own the quality of those AI experiences and the relationship with the clients who use them. This is a forward-deployed role in the practical sense: you sit between our clients and our engineering team. You will review how clients actually use Troy AI and Claudia, meet with them regularly, help them build the right prompts and skills, gather their requirements for new agents, and turn everything you learn into clear, reproducible feedback that our engineers can act on. You will not ship production code; our engineering team does that. You will diagnose, prototype, configure, and communicate. What You Will Do: * Review Troy AI and Claudia usage across our client base: what clients are asking, how accurate and useful the answers are, and where the AI fails * Classify failures by root cause (retrieval, tool or MCP call, skill or prompt design, integration gap, or model limitation) and deliver specific, reproducible improvement requests to our engineering team for the MCP server, integrations, skills, and data * Meet regularly with clients to get direct feedback on their AI usage and their overall LeadCenter experience, and bring that feedback back to the team * Help clients create and refine the prompts and skills that make Troy AI agents work correctly for their firm, and flag to engineering when the platform, not the prompt, is the problem * Meet with clients to capture their Troy AI agentic requirements, then build and configure those agents on the platform and iterate until they work * Build and maintain evaluation sets, a failure taxonomy, and a weekly AI quality report so improvements are measured, not guessed * Communicate back to clients on what we changed based on their feedback and close the loop * Own metrics such as sampled-answer accuracy, agent task completion rate, weekly active AI users per client, time from kickoff to a client’s first working agent, and escalations Roughly 40% of your time will be with clients (meetings, enablement, requirements), 30% reviewing conversations and running evals, and 30% building agent configurations, prompts, and skills and writing specs for engineering. Required Experience: * 3+ years in a customer-facing technical role at a SaaS company: solutions engineering, implementation, technical customer success, conversation design, or technical product management * Hands-on experience building working AI agents or assistants with the Claude or OpenAI APIs, custom GPTs or Claude Projects, or agent and automation tools such as n8n, Make, or Zapier * Strong prompt engineering skills: system prompts, tool-use instructions, structured outputs, few-shot examples, and evals. You can look at a bad LLM answer and explain why it failed * Working understanding of tool calling, MCP (Model Context Protocol), REST APIs, JSON, and webhooks Ability to pull and analyze logs and transcripts with SQL or Python; using Claude Code, Cursor, or similar AI coding tools to do it is expected and welcome * Experience writing clear bug reports and specs that engineers can act on without a follow-up meeting Excellent spoken and written English, and comfort leading video calls with financial advisors and firm owners Available during US Central business hours (at least 6 hours of daily overlap) * Independent and organized: you can run a client feedback program without constant direction Nice to Have: * Experience in financial services, wealth management, or the RIA industry, including awareness of SEC and FINRA compliance considerations * Experience with LLM observability and eval tools such as Langfuse, Braintrust, LangSmith, or PromptLayer Experience building MCP servers or tools, or integrating LLMs with CRMs and business systems * Experience with CRM platforms, especially those used by financial advisors * Conversation design or chatbot QA background Product management experience Who We Are Looking For: We are looking for someone practical, technical, and client-facing who already builds with AI every day. You should be equally comfortable reading 50 conversation transcripts to find a pattern, sitting on a call with an advisor to understand what they actually need, writing the prompt or skill that fixes it, and writing the ticket for engineering when it is not fixable at the prompt layer. Please do not apply if you only want to write code and not talk to clients, if you only want to talk to clients and cannot diagnose why an AI answer failed, or if you cannot work US Central hours. Engagement: fully remote in a US-compatible time zone (US, Canada, or Latin America). Compensation depending on experience and location. Hiring Process: 1- Loom video (1 minute) plus a link to one agent or automation you built 2- Work sample: we send you a set of anonymized Troy AI and Claudia conversations. You classify each failure, write the prompt or skill fix where one applies, and write one engineering ticket for a platform gap 3- 45-minute call: a short mock advisor session, then a technical discussion with our engineering team 4- Offer To Apply: Record a 1-minute video using Loom.com introducing yourself and explaining your relevant experience. Include a link to or a short demo of one AI agent, assistant, or automation you built, and tell us in one sentence what it does and what broke the first time you shipped it. We are looking for someone who can start quickly and become the long-term owner of AI quality and client success for Troy AI at LeadCenter.AI.