AURA 300

What we believe

Ambition is the point

We’re not building a nice product in a niche. We’re building the layer that powers how millions of businesses run. We hire people who want to build something that matters at that scale.

Reality is the bar

Anyone can demo. We build voice AI that has to hold up on a noisy line with a real customer waiting on the other end. Reliability in production is the standard here — not a polished pitch.

Ownership, not tasks

We’re deliberately small, and we keep it that way. You’ll own a whole surface area end to end, and the decisions that come with it. No hiding, no hand-holding, real impact.

Speed compounds

We ship, learn, and ship again. The best people here don’t wait for permission — they see what needs doing and do it. A month at Aura is worth six somewhere else.

One team, four markets, no borders

Australia, Ireland, the UK and the US run on one platform. We’re remote-first, timezone-fluid, and we treat local nuance as a design problem we solve once. Talent anywhere, excuses nowhere.

We take the work seriously, not ourselves

We’re ambitious and human. We celebrate wins, we’re honest about what’s broken, and we genuinely like building this together.

Open roles

Voice AI Engineer

India · Full-time

Build and maintain the production voice agents behind Aura — real-time speech pipelines, telephony and booking integrations, and the testing that keeps them dependable on live calls.

About the job

The Role

We are looking for a Voice AI Engineer to build reliable, scalable, and highly customized conversational AI solutions.

You will contribute to the platform daily, improving our voice agents, booking integrations, testing infrastructure, and production reliability. This role is ideal for an engineer who enjoys combining AI models with deterministic software engineering to create systems businesses can trust.

What You’ll Do

Build and maintain production voice AI agents using LiveKit Agents and Retell AI.
Develop real-time speech pipelines using Deepgram, Cartesia, OpenAI, and other model providers.
Improve interruption handling, barge-in behavior, end-of-call detection, latency, transcription accuracy, and noisy-audio performance.
Design multi-step conversational workflows for booking, rescheduling, cancellation, customer lookup, and human handoff.
Integrate agents with salon CRMs, booking platforms, payment systems, SMS, WhatsApp, and internal APIs.
Build secure tool-calling workflows with explicit confirmation, idempotency, timeout handling, and fail-closed behavior.
Create and maintain SIP and telephony integrations using LiveKit SIP and Twilio.
Develop tenant-specific prompts, tools, policies, knowledge bases, and agent configurations.
Implement RAG and dynamic knowledge retrieval for services, pricing, policies, staff, and location-specific information.
Build automated batch-testing and regression-testing systems for prompts, tools, transcripts, and complete conversations.
Create deterministic and model-based evaluations covering safety, latency, accuracy, tool usage, and conversation quality.
Run load tests, audio replay tests, failure drills, canary calls, and Retell-to-LiveKit parity testing.
Investigate production failures using correlated call, conversation, booking, and deployment telemetry.
Improve deployment pipelines, release gates, monitoring, rollback procedures, and service-level objectives.
Participate in code reviews and contribute production-quality improvements on a daily basis.

Our Technical Environment

Our current infrastructure includes:

TypeScript, Node.js, Python, FastAPI, Zod, Pydantic
LiveKit Agents and Retell AI
Deepgram, Cartesia, OpenAI, and local models through Ollama
Twilio Voice, SIP, SMS, and WhatsApp integrations
Supabase, PostgreSQL, Edge Functions, and tenant-aware APIs
Docker, AWS EC2, ECR, RDS, EBS, CloudWatch, and Secrets Manager
GitHub Actions with immutable deployments, pre-production gates, production canaries, and rollback controls
Vitest, Pytest, YAML/JSON scenario suites, transcript replay, audio testing, and custom evaluation tooling
Observability for latency, mutation errors, safety violations, handoff failures, and booking reconciliation

What We’re Looking For

Strong software engineering experience with TypeScript/Node.js or Python.
Experience building production AI agents, conversational systems, or real-time communication applications.
Strong understanding of APIs, webhooks, asynchronous systems, databases, and distributed failure modes.
Experience creating robust automated tests and debugging complex production behavior.
Ability to distinguish probabilistic model behavior from operations that require deterministic guarantees.
Familiarity with Docker, CI/CD, cloud infrastructure, monitoring, and secure secret management.
Product judgment and attention to customer experience, especially in latency-sensitive voice interactions.
Clear written communication and the ability to work independently in a fast-moving environment.

Nice to Have

Hands-on experience with LiveKit, Retell AI, Deepgram, Cartesia, or Twilio.
Experience with RAG, embeddings, retrieval quality, prompt design, and model evaluation.
Experience integrating scheduling, CRM, payment, or messaging platforms.
Knowledge of SIP, WebRTC, telephony routing, codecs, voice activity detection, and interruption handling.
Experience building multi-tenant SaaS systems.
Familiarity with Supabase and PostgreSQL.
Experience designing evaluation datasets, LLM-as-a-judge systems, golden suites, and A/B tests.
Understanding of privacy, consent, call recording, and responsible AI requirements across different regions.

Success in This Role

You will help Aura 300 deliver voice agents that answer quickly, follow business rules accurately, make safe booking changes, recover cleanly from failures, and remain dependable as call volume and customer complexity grow.

Apply

Don’t see your role?

If you’re exceptional and you believe you belong here, tell us why. We read every open application — and the best people rarely apply to a job that already exists.

Send an open application