Senior Voice AI Engineer — HIPAA-Compliant AI Receptionist Platform

Overview

Company: AISpire360
Role: Senior Voice AI Engineer (Independent Contractor, Full-Time Engagement)
Location: Remote (India-based preferred; must overlap 4–6 hours daily with London + US Eastern)
Engagement: Full-time independent contractor — long-term
Compensation: Competitive, paid in USD, monthly. Senior staff-level rates.

About AISpire360
We are building the operating system for medical practices in the US. Our flagship product is ARIA, an AI receptionist that handles inbound and outbound calls, intake, scheduling, eligibility, prior authorization workflows, and after-hours coverage for medical practices. We are in active production deployment with our first practice — a spine surgery practice at the Hospital for Special Surgery in New York. Multiple specialty practices are in our deployment pipeline.
The platform is built to be specialty-agnostic. The voice agent stack is the universal foundation. Knowledge layers (spine surgery, primary care, behavioral health, dermatology, cardiology, pediatrics, etc.) plug in as configuration. One platform, many specialties.
We are venture-backed, founder-led, and shipping. The technical decisions you make here will be in production within weeks.

What You'll Own
You are the technical anchor for the voice agent. You will be responsible for:
Real-time voice system

Voice agent stack on ElevenLabs Conversational AI (and evaluation of Retell, Vapi, LiveKit alternatives as we scale)
STT and TTS configuration, voice tuning, language and accent handling
Turn-taking, barge-in, end-of-speech detection, latency budgets
Driving end-to-end voice latency below 700ms (we currently target 600ms)
Quality vs. cost tradeoffs across voice tier selection (Premium / Turbo / Standard)
Silence-detection optimization for outbound calls (95% silence-discount realization)

HIPAA + PHI infrastructure

BAA-compliant hosting architecture (AWS-based, BAA in place)
PHI detection and redaction pipeline (AWS Comprehend Medical + custom pre-filters)
Encryption in transit and at rest, audit logging, 7-year retention compliance
Access controls, RBAC, kill-switch infrastructure
HIPAA Security Rule compliance for technical safeguards

Agent orchestration and tools

LLM integration (GPT-4o, Claude Sonnet, with model routing for cost optimization)
Tool calling for scheduling, eligibility checks, prior auth lookups, calendar booking, task creation
Multi-turn conversation state management
Multi-topic call handling and structured task generation
Safety-critical keyword detection and escalation routing

Integrations

EHR integration (Epic, Athena, eClinicalWorks — at least one in V1)
Insurance eligibility APIs (Availity, Change Healthcare, payer-specific)
Calendar integration (Google Calendar, Microsoft 365, Cal.com)
SMS and email gateways (Twilio, SendGrid)
Practice-specific webhooks for task manager surfacing

Multi-tenant platform foundations

Per-practice configuration loading (specialty profile + practice profile)
Per-tenant isolation, data segregation
Specialty-specific intake schemas, triage rules, safety keywords
Production observability (tracing, metrics, latency monitoring per practice)

Required Experience
You should be able to demonstrate, in a live walkthrough with code in hand, that you have shipped production voice AI systems. Specifically:

3+ years of production voice/conversational AI experience. Not chatbots — actual real-time voice agents handling live phone calls. ElevenLabs, Retell, Vapi, LiveKit, OpenAI Realtime, Deepgram, or equivalent.
Deep familiarity with at least 2 of: ElevenLabs Conversational AI, Retell AI, Vapi, LiveKit Agents, OpenAI Realtime API. Production deployment experience required, not demo experience.
HIPAA and PHI handling experience in production. Not theoretical. You have signed BAAs, deployed in HIPAA-compliant environments, and built PHI redaction pipelines.
AWS infrastructure expertise. Specifically: ECS or Fargate, Lambda, S3 with encryption, KMS, CloudWatch, IAM, VPC, BAA-eligible service selection.
AWS Comprehend Medical or equivalent PHI detection experience.
Twilio or equivalent telephony provider experience. SIP trunking, inbound/outbound call flows, recording and transcription pipelines.
Latency optimization in real-time systems. You have hit sub-700ms voice latency in production and can explain the trade-offs you made.
LLM production experience. GPT-4o, Claude, or equivalent. Prompt engineering, tool calling, structured outputs, model routing for cost.
Python or TypeScript fluency. Production code, not scripts.
API integration experience. REST, webhooks, OAuth, async patterns. Bonus if you've integrated with EHRs (Epic FHIR, Athena, eClinicalWorks).

Nice to Have

Healthcare domain experience (EHRs, clinical workflows, medical terminology)
Prior auth or eligibility verification workflow experience
Multi-tenant SaaS architecture experience
SOC 2 audit experience
Experience with knowledge graphs (PostgreSQL + pgvector, Neo4j, or similar)
Open-source contributions to LangChain, LiveKit, Pipecat, or similar voice/agent frameworks
Demonstrated ability to ship to production fast (not perfectionism — pragmatic engineering)

What This Engagement Looks Like

Independent contractor. You invoice us monthly. We pay in USD via wire transfer or Wise.
Full-time engagement. This is your primary work. We expect ~40 hours/week of focused output.
Long-term. We are not looking for a 3-month engagement. Expect this to run 12+ months.
You sign a BAA with us. This is non-negotiable due to the nature of the work.
You are eligible for future equity grants as we formalize the team structure.
Direct working relationship with our CTO (London-based) and founders (NYC and India).
Time-zone overlap required: at least 4–6 hours of daily overlap with both London and US Eastern. Indian Standard Time evenings (6 PM IST onward) work well.

How We'll Evaluate You
Our hiring process is fast and substantive. No take-homes. No multi-round bureaucracy.

Initial screen (30 min): Background, what you've shipped, compensation expectations, BAA-readiness.
Technical deep-dive (90 min): Walk us through a production voice AI system you've shipped. Code in hand. Architecture diagrams. Real production numbers — latency, cost per call, scale. We will ask follow-up questions about specific decisions you made.
System design (90 min, live): We give you a design problem — a HIPAA-compliant voice agent for a medical specialty we don't yet support. You design end-to-end on a whiteboard / Excalidraw. We probe trade-offs.
Founders + CTO conversation (60 min): Mutual fit. You ask us hard questions. We answer honestly.

Total: 4–5 hours of your time. We'll move from first conversation to offer in 7–10 days for the right candidate.

What We Will NOT Accept

Take-home assignments produced by other AI tools. We will ask you to walk through code live; if you can't explain decisions you didn't make, the interview ends.
Resume claims you can't substantiate. We verify employment via EPFO/UAN where applicable.
"I've done LLM chatbots" experience without production voice agent experience. They are not the same skill.
Engagement structures that conflict with this being your primary work.

How to Apply
Send a brief message with:

A link to a production voice AI system you've shipped (or, if NDA'd, a description of the architecture and your specific contribution).
Two or three sentences on why this role fits you specifically — not a generic cover letter.
Your current compensation expectation in USD per month.
Your earliest start date and current time-zone availability.

We read every message. We respond within 5 business days, including with "no" answers. We don't ghost.