🧠 Contract Position: AI Voice Agent Engineer (Long-Form Calls, Structured FSM) We are building a controlled voice agent that conducts 45–60 minute structured interviews in a healthcare/legal context. This isn’t a support bot, sales bot, or short agent — it must hold long conversations with strict branching logic and no improvisation. 🛠️ Tech we need: You should be comfortable with at least two of the following: Vapi or Retell (required) Twilio Voice / SIP / WebRTC Deepgram / Whisper streaming ASR ElevenLabs / PlayHT / custom TTS LLM orchestration with strict guardrails (GPT/Claude) Finite State Machines (XState, custom FSM, or similar) 🎯 Your responsibilities: Build a human-sounding agent that follows a deterministic script Implement reliable barge-in and low latency Store structured data + transcript as JSON output Do NOT let LLM “freestyle” Keep persona professional and legally safe (no diagnosis/advice) 💼 About the project: This agent conducts clinical-style interviews used in immigration hardship evaluations. A licensed clinician reviews the output — the AI never diagnoses. We already have full interview architecture + persona rules + JSON schema. We just need someone who can build the voice + flow + data pipeline. 💰 Engagement: Paid project-based MVP (2–4 weeks) If we work well together → ongoing contract (high-impact roadmap) We welcome talent from anywhere 📎 To be considered, please DM with: Links to 1–2 voice agents you’ve built (Retell, Vapi, or Twilio required) What stack you prefer for long calls Whether you’ve implemented barge-in before Project rate range 💬 We are moving quickly and will choose one engineer within the week. If you’ve built long-form voice agents or worked with telephony + LLM orchestration, we’d love to meet you. Send me a DM or hit me up via email at hamsaomar211@gmail.com