Looking for a founding engineer building real-time AI authenticity tooling

We’re building a real-time authenticity layer for the AI era.

The problem: interviews, assessments, and even live video are becoming increasingly synthetic. Deepfakes, copilots, teleprompters — it’s getting hard to tell what’s real in real time.

We’re building a lightweight detection system focused on:

real-time video + audio analysis

behavioral signals (latency, eye patterns, response dynamics)

practical deployment inside hiring workflows

This is not a research paper project — we’re building a working MVP now.

Stack direction:

WebRTC ingestion

Python backend (FastAPI)

real-time inference + streaming signals

lightweight front-end surface

We’re looking for a founding-level builder who:

likes hard real-time problems

is comfortable shipping rough MVPs

can work async with a small team

This is part-time initially (~40–60 hrs total to start), paid + equity.

If you’ve built real-time systems, WebRTC tooling, or weird low-latency AI stuff — would love to talk.

Happy to share more context with anyone curious.

2 comments

[ 2.7 ms ] story [ 18.5 ms ] thread

turbiakjohn 4mo ago ↗

This is a really interesting space to be working in. The whole “authenticity layer for the AI era” feels very real right now, especially with how common copilots, teleprompters, and deepfakes are becoming in interviews.

I like that you’re focused on real-time signals instead of just building another research-heavy detection model. The behavioral angle (latency, eye movement, response dynamics) sounds especially compelling if done right.

Would love to hear more about what you’ve built so far.

Happy to chat

Darshan009 4mo ago ↗

This is a genuinely interesting problem. Real time authenticity is going to matter a lot as synthetic signals get better.

I have experience building low latency backend and AI inference systems in Python and C++, including streaming pipelines and optimized model serving at scale. I have worked on sub 100ms inference systems and tuned LLM pipelines for throughput and reliability in production.

Very comfortable shipping MVPs fast, working async, and handling real time performance tradeoffs. Would love to connect and learn more.

darshangvaghasiya0@gmail.com