AI 体验工程师（对话式 AI）

About the Role:

An AI Experience Engineer owns the full lifecycle of the journeys our customers have with Upgrade's AI agent on chat, voice and email: creation, optimization, edge case handling and containment growth. They build journeys in Sierra's no-code tooling, partner with Upgrade's product groups to bring their servicing flows to the agent, and are accountable for whether the agent resolves the customer's issue safely, accurately and without a human.

We are hiring an AI Experience Engineer to own Upgrade's customer-facing AI agent. You will sit between engineering, product and customer operations. You will turn servicing workflows, such as making a payment, checking loan details, updating an address or enabling autopay, into production journeys and keep improving them after launch. You will build and own the journeys end to end, from first draft to steady state. The work is part builder, part analyst: design the flow, handle the edge cases, read the transcripts, and keep raising containment. Containment, CSAT, transfer quality and hallucination rate are your metrics.

Within your first year, you will own a portfolio of journeys across products (Personal Loans, Card, BNPL, HELOC, Deposits) and help set the standards for how Upgrade builds, tests and ships AI experiences.

Our Core Technologies

Sierra Agent Studio (no-code journeys), Sierra simulations and conversation analytics, GraphQL and REST APIs (as journey tools), Genesys (contact center handoff), SumoLogic, Grafana, SQL, Javascript/Typescript, Claude and other LLM tools.

What You'll Do

• Journey creation: Design and build Sierra journeys in Agent Studio for chat, voice and email, from discovery with operations through production rollout

• Optimization: Review transcripts, recordings and analytics every week, find where conversations fail or transfer, and ship fixes to journey logic, instructions and knowledge

• Edge case handling: Map the unhappy paths (failed payments, mismatched identity, delinquent accounts, unsupported requests) and make sure each ends in a correct answer or a clean handoff

• Containment: Own containment and CSAT targets per journey, run staged rollouts (1% → 100%), and prioritize the work that moves those numbers most

• Product partnership: Work with product groups (Personal Loans, Card, BNPL, Deposits) to scope new journeys, keep journeys current as products change, and define the APIs the agent needs

• Safety and compliance: Apply tiered authentication and step-up verification, verbatim compliance language and guardrails against hallucination and off-policy answers

• Handoffs: Design transfers to human agents in Genesys with accurate summaries and auth context

• Testing: Build simulation and regression suites so every journey change ships with evidence it did not break existing behavior

• Monitoring: Watch latency, errors and drop-offs in SumoLogic and Grafana, and own the alerts

• Voice tuning: Latency, filler phrases, interruptions, personality and tone

• Sierra relationship: Be the technical owner on our side, escalate platform issues and shape roadmap asks

What We Look For:

• 3–5 years of Frontend Software Engineering experience, or equivalent experience in Solutions Engineering, Conversational AI or Technical Product roles

• Working knowledge of how LLM agents behave: instructions, tool calls, knowledge retrieval, and where each one fails

• Experience building evaluations for non-deterministic systems: test sets, simulations, LLM-as-judge, regression gates

• Enough technical depth to read API specs, work with backend engineers on REST and GraphQL integrations, and reason about authorization and permission scopes

• Working knowledge of JavaScript/TypeScript, with the ability to build and maintain coded journeys and integrate APIs within conversational AI platforms.

• Product and customer mindset: you understand ops procedures, you read transcripts, care about tone, and measure success by whether the customer's problem got solved

• Ownership in production: rollouts, monitoring, log analysis and incident response

• Clear writing and the ability to work with non-engineers in operations, compliance and legal

Nice To Have:

• Hands-on experience with Sierra Agent Studio, or a comparable no-code agent platform (Decagon, Ada, Kore.ai, Google CCAI, Amazon Lex/Connect)

• Voice AI experience: speech-to-text, text-to-speech, latency tuning, barge-in handling

• Contact center systems such as Genesys, and agent handoff design

• Background in fintech or another regulated industry (UDAAP, FDCPA, TCPA, PCI, SOC 2)

• Conversation design or UX writing experience

What We Offer You:

• Competitive salary ($140K-$180K CAN) and stock option plan

• Paid coverage of medical, dental and vision insurance

• Flexible PTO

• Opportunities for professional growth and development
• Paid parental leave

• Health & wellness initiatives