Job Summary
The AI Support Operations Lead is the technical architect of Mercans’ AI-agentic support model. This person designs, builds, deploys, and continuously improves the AI systems that handle L1 triage, automated resolution, knowledge base management, self-service, and support analytics. They work at the intersection of AI/ML engineering and support operations, working closely with the L3 team to convert root cause findings into AI resolution patterns, creating a virtuous cycle where human expertise continuously amplifies AI capability. This is a production-focused, outcomes-driven position — not a research role.
Duties & Responsibilities
AI System Design & Deployment
Design and build the AI agentic support system integrated with YouTrack: automated categorization, priority assignment, routing, and auto-resolution where confidence thresholds are met
Develop the client-facing self-service AI interface for common queries (payslip requests, form downloads, status checks) without ticket creation
Monitor AI agent performance daily; retrain models, adjust confidence thresholds, and expand resolution coverage
Manage AI-related vendor tools and APIs; evaluate new AI capabilities and recommend adoption
Knowledge Base & Training Data
Build and maintain the knowledge base powering AI agents: structured resolution patterns, FAQ databases, client-specific runbooks, and compliance reference materials
Work with L3 Engineers to convert their root cause findings and resolution patterns into AI training data and automated resolutions
Establish the AI feedback loop: ensure 100% of human-resolved tickets are captured as potential AI training data
Analytics & Automation
Implement real-time support analytics: ticket volumes, SLA health, AI triage accuracy, auto-resolution rates, bounce rates, recurring issue clusters, deployment impact
Automate operational processes: SLA alerting, on-call notifications, status updates, ticket aging alerts, and weekly reporting
Provide data-driven recommendations to the Head of Support on workload balancing, staffing needs, and process improvements
Collaboration
Collaborate with the Customer Sentiment Pod on the proprietary AI agent roadmap
Work with Support Engineers to capture resolution patterns from every human-resolved ticket
Skills & Qualifications
Required Competencies
Proficiency with LLM architectures (Claude, GPT, custom models), NLP techniques, prompt engineering, RAG, and agent framework design
Strong programming skills (Python, JavaScript/TypeScript)
Experience with API integrations, database querying, and ITSM platform APIs (YouTrack, Jira)
Data visualization capabilities (Grafana, Tableau, or equivalent)
Understands support workflows, SLA management, and ticket lifecycle
Able to translate operational pain points into automation solutions
Experience & Education
Bachelor’s degree in Computer Science, Data Science, AI/ML, or related technical field. Advanced degree preferred
Minimum 3 years in AI/ML engineering, support automation, or data engineering within a SaaS environment
Direct experience building AI-powered support tools (chatbots, NLP categorization, auto-resolution systems) required
Experience fine-tuning models on domain-specific data
SMART Goals
AI Auto-Triage Deployment
Specific: Deploy AI auto-categorization and routing for all incoming support tickets across all clients
Measurable: Percentage of tickets auto-categorized with 85%+ accuracy
Achievable: Using LLM-based categorization trained on historical ticket data and client configurations
Relevant: Eliminates manual queue management and enables the dedicated account model
Time-bound: 100% coverage with 85%+ accuracy within 60 days of hire
Auto-Resolution Rate
Specific: Achieve 20% auto-resolution rate on L1 tickets
Measurable: Percentage of tickets fully resolved by AI without human intervention
Achievable: By building resolution patterns from historical data and L3 documentation
Relevant: Core metric for the lean, AI-first operating model
Time-bound: 20% auto-resolution within 90 days of hire
Knowledge Base Foundation
Specific: Build structured KB entries covering the top 50 recurring support issues
Measurable: Number of KB articles indexed and searchable by AI
Achievable: By analyzing historical ticket patterns and working with Support Engineers and L3 team
Relevant: Foundation for AI auto-resolution and client self-service
Time-bound: 50 KB articles within 75 days of hire
Analytics Dashboard
Specific: Launch a real-time analytics dashboard covering all core KPIs
Measurable: Dashboard operational with auto-refreshing data, reviewed in weekly meetings
Achievable: Using existing ticket data from YouTrack and building automated data pipelines
Relevant: Enables data-driven management of the support function
Time-bound: Dashboard live within 30 days of hire
AI-L3 Feedback Loop
Specific: Establish the process for converting L3 root cause closures into AI training data
Measurable: Percentage of L3 closures fed into the AI learning pipeline
Achievable: Working with L3 engineers to structure their findings into machine-readable patterns
Relevant: Creates the flywheel where human expertise amplifies AI capability
Time-bound: 100% of L3 closures fed to AI pipeline within 45 days of hire
Self-Service Pilot
Specific: Launch an AI-powered self-service pilot for the top 3 query types (e.g., payslip visibility, form requests, status checks)
Measurable: Pilot live and handling queries for at least 2 clients
Achievable: Building on the KB foundation and AI triage infrastructure
Relevant: Reduces ticket creation at the source
Time-bound: Pilot live for 2+ clients within 90 days of hire