Hemut / AI voice engineering
Voice agents for freight calls, with work spanning the conversation, the backend, and how both reach production.
Shipment reference muted / Original timing
Agent: Thanks for calling Sunview Logistics. How can I help you today? Caller: Hi, can I check up on a load status? Agent: Yeah, I can do that. What's your load or reference number? Caller: [Load reference muted.] Agent: Alright, one moment. Right, [load reference muted] is delivered.
Automatic transcript, edited to mark redactions. 19-second excerpt. Shipment reference muted; original pace and pauses preserved. The source is a 48 kHz Opus recording.
Telephony input
Speech to text
Context and tools
Text to speech
I did the voice engineering and research and wrote nearly all of the Converse documentation. My work in Hevox also included Alembic database migrations and deployment work, including SHA-pinned ECS releases. Document synchronization was another part of my backend work at Hemut.
I chose LiveKit after my own latency comparisons, including work with Meiborg Brothers. The decision came from the agents and calls I was working on, rather than a claim that one voice platform is always the fastest.
I caught end-of-utterance and transcription timings being added together even though they overlapped. I also found two speech-recognition streams running at the same time; they had been running for a few days. These were separate findings about what the runtime was actually doing.
Reported time-to-first-speech comparison following request parallelization, pooled HTTP connections, and end-of-turn tuning. The sampling window and percentile are unspecified.
Original resume reports a 63% reduction after moving STT, LLM, and TTS onto LiveKit's native Inference path and improving prompt caching. No matched billing window is attached to this comparison.
Hit rate reported in the original resume's voice-cost result. Its denominator and measurement window are unspecified; it is not interchangeable with a separate invoice's cached-token share.
Original resume compares the earlier Drive walk with the document-sync result. These are reported timings without an attached workload, sample size, or matched run.
Cost reduction reported alongside the document-sync timing in the original resume. The underlying billing window and workload are not specified.
Voice engineering, research, documentation, migrations, and deployment work by DongYeop Lee at Hemut.
Voice agents for freight calls, with work spanning the conversation, the backend, and how both reach production.
A logistics copilot that finds the tools a question needs and coordinates answers across them.
A speech-recognition benchmark built around freight calls, including a scoring correction that changed my model comparison.
