SENIOR AI / TWILIO ENGINEER — REAL-TIME VOICE + SMS ENGINE
Budget: $1,500 – $3,000 USD
(If this scares you, don’t apply.)
We’re building a production-grade AI communication engine — real-time voice, SMS, WhatsApp, memory, error recovery, the works.
If you’ve only done “Twilio scripts,” stop reading now. This isn’t that.
We need someone who thinks in architecture, moves fast, and builds systems that scale.
⸻
Real-Time Voice AI (Twilio Streams)
• Live audio streaming to LLMs
• Whisper messages + escalation logic
• Call routing + fallback
• Error recovery + retry flow
• Production-grade, low-latency design
⸻
SMS + WhatsApp AI Engine
• Intent router
• Conversation state machine
• Memory layer (short + long-term)
• Booking / order flows
• Automated follow-ups
• Webhooks + retry logic
• Multi-channel compliance (A2P, WhatsApp templates)
⸻
Reliability Layer
• Freeze detection
• Model fallback
• Queueing
• Logging + monitoring
• High-volume throughput (100k+ msgs/day)
⸻
Ideal Candidate Must Have
:heavy_check_mark: AI / LLM engineering (OpenAI / Anthropic / Mistral)
:heavy_check_mark: Structured output, function calling
:heavy_check_mark: Prompt engineering at scale
:heavy_check_mark: Memory design
:heavy_check_mark: Intent classification
:heavy_check_mark: Backend engineering
• FastAPI or Node
• Redis / queues
• Docker
• Event-driven architecture
• Scalable systems
• Testing + monitoring
⸻
:x: DO NOT APPLY IF:
• You only know Twilio Studio / TwiML
• You’ve never built a state machine
• You rely on ChatGPT to write your code
• You can’t explain queues, retries, fallbacks
• You’ve never touched real-time voice streaming
• You’ve never shipped an AI product end-to-end
:white_check_mark: APPLY IF:
• You’ve built REAL AI communication systems
• You architect systems, not “scripts”
• You move fast and communicate clearly
• You can own this engine from day one
• You understand what production means
⸻
First 14 Days Deliverables
Week 1
• SMS + WhatsApp AI MVP
• State machine + intent router
• Memory layer v1
• Webhook endpoints + test flows
Week 2
• Twilio Voice prototype
• Real-time audio streaming
• AI answering calls + basic logic
• Error recovery + fallback
If these timelines scare you — don’t apply.
⸻
Who This Role Is For
A senior engineer who has actually built AI communication systems from scratch.
Someone who thinks in architecture, not scripts.
A problem solver who moves fast, solves problems, and communicates cleanly.
⸻
How to Apply (Non-Negotiable Requirements)
Send:
3 links/screenshots of past Twilio Voice / real-time AI work
1–2 paragraphs describing an AI state machine you built
Exact plan for your first 7 days
Availability + timezone
If you cannot provide these, do not apply.
⸻
Skills Required
Twilio • Realtime Audio • Node/FastAPI • Redis • LLMs • Prompting • API design • Scalable systems
We’re building a production-grade AI communication engine — real-time voice, SMS, WhatsApp, memory, error recovery, the works.
If you’ve only done “Twilio scripts,” stop reading now. This isn’t that.
We need someone who thinks in architecture, moves fast, and builds systems that scale.
⸻
Real-Time Voice AI (Twilio Streams)
• Live audio streaming to LLMs
• Whisper messages + escalation logic
• Call routing + fallback
• Error recovery + retry flow
• Production-grade, low-latency design
⸻
SMS + WhatsApp AI Engine
• Intent router
• Conversation state machine
• Memory layer (short + long-term)
• Booking / order flows
• Automated follow-ups
• Webhooks + retry logic
• Multi-channel compliance (A2P, WhatsApp templates)
⸻
Reliability Layer
• Freeze detection
• Model fallback
• Queueing
• Logging + monitoring
• High-volume throughput (100k+ msgs/day)
⸻
Ideal Candidate Must Have
:heavy_check_mark: AI / LLM engineering (OpenAI / Anthropic / Mistral)
:heavy_check_mark: Structured output, function calling
:heavy_check_mark: Prompt engineering at scale
:heavy_check_mark: Memory design
:heavy_check_mark: Intent classification
:heavy_check_mark: Backend engineering
• FastAPI or Node
• Redis / queues
• Docker
• Event-driven architecture
• Scalable systems
• Testing + monitoring
⸻
:x: DO NOT APPLY IF:
• You only know Twilio Studio / TwiML
• You’ve never built a state machine
• You rely on ChatGPT to write your code
• You can’t explain queues, retries, fallbacks
• You’ve never touched real-time voice streaming
• You’ve never shipped an AI product end-to-end
:white_check_mark: APPLY IF:
• You’ve built REAL AI communication systems
• You architect systems, not “scripts”
• You move fast and communicate clearly
• You can own this engine from day one
• You understand what production means
⸻
First 14 Days Deliverables
Week 1
• SMS + WhatsApp AI MVP
• State machine + intent router
• Memory layer v1
• Webhook endpoints + test flows
Week 2
• Twilio Voice prototype
• Real-time audio streaming
• AI answering calls + basic logic
• Error recovery + fallback
If these timelines scare you — don’t apply.
⸻
Who This Role Is For
A senior engineer who has actually built AI communication systems from scratch.
Someone who thinks in architecture, not scripts.
A problem solver who moves fast, solves problems, and communicates cleanly.
⸻
How to Apply (Non-Negotiable Requirements)
Send:
3 links/screenshots of past Twilio Voice / real-time AI work
1–2 paragraphs describing an AI state machine you built
Exact plan for your first 7 days
Availability + timezone
If you cannot provide these, do not apply.
⸻
Skills Required
Twilio • Realtime Audio • Node/FastAPI • Redis • LLMs • Prompting • API design • Scalable systems