Senior AI Engineer — Conversational Voice AI

**About SmartAdvocate** SmartAdvocate is an industry\-leading legal case management platform used by 20,000\+ legal professionals across North America, from solo practitioners to large multi\-office firms. We are embedding advanced AI directly into the core of the platform to automate complex legal workflows, modernize firm operations, and generate deep document intelligence. **What you would be joining** This is not a greenfield AI experiment. SmartAdvocate already runs a suite of AI services in production: document classification and data extraction, medical chronology generation, demand letter drafting, case chat over vectorized case documents, an agent runtime, and an MCP server that lets external AI platforms operate directly against case data. Voice is the next frontier: intake voice agents that handle new\-case calls end to end, a receptionist agent that answers client questions from live case context, and outbound agents that confirm appointments, chase medical records, and keep clients updated. You would own that work. **Position overview** We are looking for a hands\-on senior engineer to build our conversational voice AI: enterprise\-grade voice agents wired into telephony that handle intake and client calls, understand the caller's case, and write structured results straight into the case file. This is a build\-and\-operate role. You will design the pipelines, ship them, and own them in production. ***Key responsibilities*** **Voice agents and telephony** * Architect, build and deploy conversational voice agents that interact with callers, leads and clients. * Integrate with VoIP/SIP and cloud communications platforms (Twilio, Plivo, Amazon Connect, LiveKit) across the topologies our customers actually use: call forwarding, conferencing an SA number onto a live call, and full number porting — so we work with any firm's phone system. * Handle the realities of live calls: barge\-in and turn\-taking, interruptions, poor reception, accents, background noise, and graceful degradation when confidence drops. * Optimize real\-time STT, TTS, NLU and latency budgets; select and tune voices for naturalness. * Design warm transfer and escalation — handing a call to a human with full context, and recognizing high\-value or urgent situations that should be routed immediately. **Live data extraction and write\-back** * Build real\-time pipelines that transcribe, parse and structure call conversations to populate case fields, create intake records, update case notes and trigger downstream workflows automatically. * Record and summarize every call into the case file with clear provenance for anything the agent entered. * Detect the caller's language and respond in it — Spanish is our first priority after English. **Safety, configuration and operations** * Design caller identification and authentication before any case information is disclosed. Caller ID alone is not identity. * Make behavior configurable per firm: operating hours and aft