1+ years of software development experience, including internships.
Strong engineering fundamentals with production code ownership.
Comfort with observability tools like OpenTelemetry.
Startup-ready mindset for broad ownership and fast iteration.
Ability to communicate AI behaviors and tradeoffs clearly.
Responsibilities
Build and enhance the real-time voice AI pipeline for speech-to-text and text-to-speech.
Improve model performance post-deployment and run evaluations.
Develop instrumentation, dashboards, and alerting mechanisms.
Optimize costs by increasing throughput with open-source LLMs.
Collaborate with major vendors on evaluating new voice and language models.
Deploy AI R&D features such as voice cloning and multilingual capabilities.
Implement safety measures (AI guardrails) for live customer interactions.
Benefits
Equity participation through stock options.
Fully paid health, vision, and dental insurance for employees.
Flexible PTO and hybrid working schedule.
Free lunches and parking stipend, along with team events.
Casual yet productive company culture.
Full Job Description
About the Role
We're hiring a Software Engineer (AI/ML) to work on our voice AI platform, the real-time pipeline of speech-to-text, LLM reasoning, and text-to-speech that powers every call. You'll work on latency, model quality, cost, and reliability across the full AI stack, and help shape which models and vendors we bet on. This is a deep, focused role on a small core product team, with direct mentorship from our AI lead.
What You'll Do
Build and improve the real-time voice AI pipeline: speech-to-text, conversational LLM orchestration, and text-to-speech, with hard latency targets (we hold P90 end-to-end latency under 2 seconds and are pushing lower).
Improve model performance in production, run evals for core functionality
Improve instrumentation (OTel/Braintrust), build dashboards and alerting
Optimize cost: squeezing more throughput out of open source LLMs
Work with vendors like Anthropic, Microsoft/Azure OpenAI, Deepgram, and NVIDIA - testing new voice and language models, evaluating alternatives, and handling failover when a provider has issues.
Ship AI R&D features into production: things like backchanneling, emotional expressiveness, background audio, voice cloning, and multilingual features
Implement AI guardrails so models behave correctly on live customer calls
What We're Looking For
2 -5 years of experience in software development (including relevant internships).
Strong engineering fundamentals: you write production code, instrument it, and own what you ship
Comfort with observability tooling (OpenTelemetry or similar) and making decisions from metrics
Scrappy, startup-ready mindset: small team, broad ownership, fast iteration
Clear communicator - you'll explain AI behavior and tradeoffs to product, support, and customer-facing teams
Nice to Have
Experience with real-time audio or voice AI (STT/TTS pipelines, streaming, WebRTC/networking, telephony)
Hands-on experience building with LLMs in production: prompt design, model evaluation, latency/cost optimization
Background in speech/audio ML (a founder or early builder of a voice product is a great signal - that's how our current Head of AI got here)
Azure cloud experience
Why Mia
Equity participation through stock options.
Fully paid health, vision & dental insurance for employees.
Flexible PTO and hybrid schedule (3 days / week in downtown Austin).
Free lunches, parking stipend provided, team events, and a casual, but get things done culture.
Compensation
Compensation for this position will be competitive with current market rates and reflective of the candidate's experience, skills and qualifications.
