VoiceAI builds the real-time voice layer behind Trusting Social's AI agents, for both outbound and inbound calls. We own the cascade pipeline that connects the caller to the agent (ASR → LLM → TTS): real-time speech understanding (ASR), real-time speech generation (TTS), the real-time orchestration between speech and the LLM script, and the interaction behaviour on top, including turn-taking, latency, and answering machine detection. Our platform powers Kompato (AI collections for US fintech lenders) and Sophia (financial services sales in Vietnam and the Philippines), working closely with the Collin and Sophia DS and product teams, who own the LLM dialogue scripts, and the engineering team, who owns telephony and inbound routing and connects to our layer through the same interface for inbound and outbound calls.

About the role

You will build and operate the real-time interaction layer that makes our agents sound and behave like great human callers: fast, natural, and reliable under production load.

What you'll do

• Own the real-time orchestration of the cascade pipeline (ASR → LLM → TTS): streaming transcripts into the LLM script, streaming its output into speech generation, and managing state and timing across components
• Build and maintain the real-time voice components: real-time speech understanding, real-time speech generation, voice activity and end-of-turn detection, barge-in and interruption handling
• Drive end-to-end latency down across the interaction layer, from audio in to first audio out
• Integrate and switch between ASR/TTS vendors and self-hosted speech models behind clean, well-tested interfaces
• Deploy and scale self-hosted speech models on GPU (batching, streaming inference, cost per call)
• Own reliability of the interaction layer: monitoring, alerting, load testing, and incident response with Datadog
• Work with Data Scientists to ship model improvements (e.g. AMD, TTS post-training) into production safely via A/B tests and staged rollouts
• Collaborate with the Collin and Sophia DS and product teams on the interface between speech and the LLM script, and with engineering on telephony and inbound routing, maintaining a single interface for inbound and outbound calls

What we're looking for

• Hands-on with a full-stack mindset: you own the whole lifecycle, from modelling and engineering to deployment and production monitoring
• AI-native way of working: you use AI tools (coding agents, LLM assistants) every day to build, analyze, and ship faster
• Real-time mindset, especially for voice and speech: you design for streaming and low latency by default, think in milliseconds, and care about how every delay and glitch sounds to the caller
• 3+ years of backend or ML engineering, with production systems that handle real-time or streaming workloads
• Strong Python; experience with async/concurrent programming and WebSocket or WebRTC-based streaming
• Hands-on experience with real-time voice frameworks (LiveKit, Pipecat, or similar) or equivalent media pipelines
• Experience integrating ASR/TTS APIs and reasoning about audio formats, codecs, buffering, and jitter
• Solid engineering practices: testing, CI/CD, containers, cloud deployment, observability

Nice to have

• Serving speech or LLM models on GPU (vLLM, Triton, TensorRT, or similar)
• Familiarity with SIP/telephony concepts
• Experience with speech-to-speech models or cascade vs. S2S architecture trade-offs
• Vietnamese, Tagalog, or US English voice product experience

What We Offer

Join our vibrant team and enjoy:

• Opportunity to work and learn from one of the best and brightest technology teams in Vietnam
• Be part of a winning team with exponential growth regionally, experience recruiting world-class talents
• Competitive compensation package, including 13th-month salary and performance bonuses
• Comprehensive health care coverage for you and your dependents
• Generous leave policies, including annual leave, sick leave, and flexible work hours
• Convenient central district 1 office location, next to a future metro station
• On-site lunch with multiple options, including vegetarian
• Grab for work allowance and fully equipped workstations
• Fun and engaging team building activities, sponsored sports clubs, and happy hour every Thursday
• Unlimited free coffee, tea, snacks, and fruit to keep you energized
• An opportunity to make a social impact by helping to democratize credit access in emerging markets

At Trusting Social, we live by ownership, integrity, and agility in execution. We believe in doing what's right, what's best, and what's innovative. If you're smart, driven, and want to make a difference in the world with the most advanced and fascinating technology, come join our team. We offer the runway to truly make an impact.

Learn more about us: https://trustingsocial.com | https://www.youtube.com/watch?v=inAEDGvOcL8&t=29s