Citrin Cooperman's Applied AI Practice designs, builds and runs AI agents inside clients' businesses. Those clients are middle-market companies, private equity sponsors and their portfolio companies, and enterprise clients our team has served for years.As a Forward Deployed Engineer, you own the technical side of an engagement. You work in a pod with a Deployment Strategist, who owns the business case, the scope and the executive relationship. You own the architecture, the code, and whether the system holds up in production.Most of our clients don't have large IT or data teams. A typical engagement starts with scattered data, an older ERP, a small IT staff, and a CFO who wants to see the controls before an agent touches the general ledger. You leave with a system in production and a client team that can run it.WHAT YOU'LL BUILDAgents for sales, marketing, service, operations, finance and accounting work, connected to clients' ERPs, accounting systems and CRMsMCP servers, agent skills and integrations that give models safe access to client systems, written so the next engagement can reuse themData pipelines that prepare client data for AI: you profile, clean, model and connect the sources before an agent depends on themEvaluation suites, monitoring and audit trails, so every agent ships with a measured accuracy baseline and every action it takes can be traced and reviewedDISCOVERY AND DESIGNRun technical discovery with your Deployment Strategist: map the client's workflow, systems and data, then turn the business problem into an architecturePick the right stack for each client: Anthropic or OpenAI models, AWS, Azure, GCP, Salesforce Agentforce, or a mixBuild proofs of concept quickly, and know what it takes to turn one into a production systemDELIVERY AND PRODUCTIONOwn components from architecture through deployment, monitoring and handoff in the client's environmentGet systems through client security reviews, access controls and data-handling requirements, including at regulated clientsTrain client teams to operate and extend what you deliverPRACTICE BUILDINGTurn what works into reference architectures, starter repos, eval templates and playbooks for the rest of the PracticeReview designs and code from other engineers on the teamSend field lessons and platform gaps back to the Practice and to our technology partnersWhat Success Looks Like in the First YearSystems in production at clients, each with a measured accuracy baseline and an audit trailReference architectures and starter repos that other engagements reuseYou meet the billable utilization target and earn strong client satisfaction scoresFive to eight years of software engineering experience, including at least one production system you can walk us through in detailClient-facing technical delivery experience in consulting, professional services or a forward deployed role; this is a requirementDeep expertise in at least one platform: Salesforce, AWS, Azure, GCP, or a frontier lab ecosystem (Anthropic or OpenAI APIs, agent SDKs and tooling)Production experience with large language models: prompt design, tool use, agents and evaluation, including why an agent failed and what you would changeFluency in Python or a strongly typed language, and working SQLExperience designing APIs, data models and integrations, beyond using themAble to explain technical trade-offs to finance leaders and executives as clearly as to engineersWilling to travel about 25% of the timePreferred QualificationsFinance and accounting operations experience: close, AP/AR, FP&A or auditPrivate equity experience: sponsor work, portfolio operations or value-creation plansDelivery in regulated industries such as financial services or healthcareMiddle-market consulting, where clients have small IT teams and fixed budgetsCertifications from Anthropic (Claude Certified Developer, Claude Certified Architect), Salesforce (Agentforce Specialist, Platform Developer I), AWS, Azure or GCPExperience with MCP, agent frameworks (Claude Agent/Messages SDK, OpenAI Agents SDK, LangGraph) or cloud data platforms (Snowflake, Databricks, Microsoft Fabric, Salesforce Data 360)