Hey everyone, Mukul here. To add a bit of technical context, the biggest headache we are actively navigating right now is balancing the audio latency with the time it takes the LLM to process the candidate's last answer and generate a relevant, probing follow-up.

We are currently leveraging Supabase and AWS to handle the backend routing and logic. Keeping the conversation feeling natural without awkward pauses is our absolute top priority. If anyone has experience shaving off milliseconds in real-time conversational AI setups, I am all ears.

Also, I know automated interviews can be a controversial topic when it comes to candidate experience. We really tried to make this feel like a helpful screening chat rather than a robotic interrogation. If you decide to test it out, I would love your brutal honesty on whether we are hitting that mark or totally missing it.

Happy to answer any questions!