I’ve been working on InterviewFlowAI, a tool that automates the first-round hiring workflow for teams that spend too much time on initial screening. It handles resume scoring, public job links, candidate applications, and full interviews conducted over phone or Google Meet.

I built this after spending years interviewing candidates as a Head of Engineering and realizing that most of the bottleneck happens before an engineer ever gets involved. Too many resumes, too many unqualified applicants, and a ton of repetitive phone screens that rarely lead anywhere.

Here’s what the system does today: • Generates a public job link so candidates can apply directly • Scores resumes based on the job description and required skills • Lets you accept or reject candidates instantly • Runs a live interview (phone or Google Meet) using an AI agent • Produces a structured scorecard, transcript, and recording

Technical details for those interested: • Uses OpenAI real-time API for conversation flow • Voice handling and telephony via Vapi • Speech → text → scoring pipelines using AssemblyAI and custom logic • Resume scoring uses embeddings + rule-based signals to reduce LLM hallucination • All interviews are stateless interactions stored securely for evaluation

Pricing is $0.50 per interview, so teams can screen at scale without high per-candidate costs.

I’d love feedback from the HN community, especially around accuracy, bias concerns, architecture, and scaling. Happy to answer any questions about design decisions or what’s still rough.