Description
We are looking for a Conversation AI Specialist who builds agentic, process-aligned dialogue systems , not scripted chatbots.
You will build the conversational intelligence layer of our advisor efficiency platform. It includes timely orchestration, intent reasoning, and LLM-powered assist systems that help advisors during complex customer interactions across chat, phone, and ecommerce. You will own the full conversation stack, from utterance to action.
Key Responsibilities
- Agentic Conversation Design: Architect and build the conversational layer of our prompt-based agentic platform , including intent detection, dialogue orchestration, and multi-step reasoning flows aligned to advisor workflows.
- LLM Integration & Prompt Engineering: Design and maintain the prompt architecture that drives advisor-assist features , including context injection, tool-calling patterns, guardrails, and output validation.
- Advisor Assist Systems: Build real-time AI tools that surface relevant knowledge, suggest responses, and guide advisors through commerce, subscription, and support interactions.
- Process Mapping: Work with operations and product teams to deeply understand advisor business processes, then translate that understanding into conversation system design.
- Evaluation & Quality: Define conversation quality metrics, build evaluation datasets from real interaction logs, and operate a continuous improvement loop for deployed systems.
- Probabilistic Fallback Design: Design graceful degradation patterns , ensuring the system handles low-confidence states, ambiguous intents, and out-of-scope inputs without advisor disruption.
- Cross-platform Delivery: Integrate conversational AI features throughout chat, phone, and ecommerce advisor surfaces , maintaining coherent behavior across channels.
Attitude & Approach We're Hiring For
Beyond technical skills, we are specifically hiring for three attitude shifts that define success in this role:
- Process-aligned composition: you build conversation flows that mirror advisor operational workflows , not abstract dialogue trees disconnected from how work actually happens.
- Probabilistic reasoning: you architect systems where intent is inferred, confidence is tracked, and ambiguity is handled gracefully rather than failing hard.
- Data-driven iteration: you measure conversation quality with data, not intuition , defining metrics, curating evaluation sets, and iterating based on evidence.
Required Qualifications
- 3–5 years in conversational AI, NLP engineering, or LLM-based product development.
- Experience building dialogue systems with real business process alignment, not just scripted flows.
- Proficiency in Python and NLP tooling
- Strong understanding of intent classification, entity extraction, and context management in stateful conversations.
- Experience deploying conversational AI to production with real user traffic.
- Ability to build for uncertainty , confidence scoring, fallback handling, human-in-the-loop blocking issue.
Preferred Qualifications
- Experience building agentic systems with tool use, function calling, or multi-step LLM orchestration. Familiarity with RAG for knowledge-grounded responses.