Upcoming Trainee Engineer @ Simform 🚀
Building, experimenting, and writing my way through tech.
AI • Web • Open Source • Challenges
If I find something interesting, I'll probably write about it. 😄
Location
Vadodara, Gujarat, India
Education
B.Tech in Computer Engineering @ Dharmsinh Desai University
Pitching under pressure is one thing. Recovering when something goes wrong mid-presentation is a completely different skill.
That's what interested me about building PitchPulse. Instead of giving speakers another generic confidence score, I wanted to help them identify a difficult moment, observe their physiological signals, reset, and rehearse that same moment again using the Ghost Take comparison.
A particularly important design decision was keeping the feedback honest: no fabricated recovery metrics, no confidence diagnosis, and no pretending simulated data is live data.
I'd love to know: What's harder for you during a high-pressure presentation — staying composed, recovering after losing your train of thought, or handling an unexpected question?
Upcoming Trainee Engineer @ Simform 🚀
Building, experimenting, and writing my way through tech.
AI • Web • Open Source • Challenges
If I find something interesting, I'll probably write about it. 😄
Location
Vadodara, Gujarat, India
Education
B.Tech in Computer Engineering @ Dharmsinh Desai University
Upcoming Trainee Engineer @ Simform 🚀
Building, experimenting, and writing my way through tech.
AI • Web • Open Source • Challenges
If I find something interesting, I'll probably write about it. 😄
Location
Vadodara, Gujarat, India
Education
B.Tech in Computer Engineering @ Dharmsinh Desai University
Upcoming Trainee Engineer @ Simform 🚀
Building, experimenting, and writing my way through tech.
AI • Web • Open Source • Challenges
If I find something interesting, I'll probably write about it. 😄
Location
Vadodara, Gujarat, India
Education
B.Tech in Computer Engineering @ Dharmsinh Desai University
Hey Dhruv! This is such a clever concept—using a "Ghost Take" mechanism inspired by racing games for public speaking recovery is brilliant, and props for keeping the data locally stored without telemetry bloat.
Quick question out of curiosity: since Electron + React setup handles the 60 FPS rendering and native bindings smoothly for local testing, how would you approach stripping this down if someone wanted to run it as a lightweight, cross-platform CLI/daemon or on lower-end edge hardware without relying on the full Electron desktop shell? Would love to hear your take on keeping that real-time canvas smoothness on a leaner stack!
The Ghost Take idea is the part that makes this more than another score-after-the-fact tool: rehearsing against your own previous pulse curve turns recovery into something you can actually practise, not just read about. I also liked the honesty guard — showing "No measurable pressure moment captured" instead of inventing a flattering metric — and pausing breathing updates while speaking, since motion would otherwise pollute the trace. One question: when the Spacebar-marked moment is a forgotten transition rather than a pricing slide, does the recovery pattern look different enough in the pulse timeline to tell those two kinds of pressure apart?
We're a place where coders share, stay up-to-date and grow their careers.
Pitching under pressure is one thing. Recovering when something goes wrong mid-presentation is a completely different skill.
That's what interested me about building PitchPulse. Instead of giving speakers another generic confidence score, I wanted to help them identify a difficult moment, observe their physiological signals, reset, and rehearse that same moment again using the Ghost Take comparison.
A particularly important design decision was keeping the feedback honest: no fabricated recovery metrics, no confidence diagnosis, and no pretending simulated data is live data.
I'd love to know: What's harder for you during a high-pressure presentation — staying composed, recovering after losing your train of thought, or handling an unexpected question?
Cool bro. I'll try it for sure. But after exams.........
Good luck 🔥
Thanks, Di.😄
No worries. Best of luck though!
You too!
Good luck @yug_vasava bro and @dj29 bro 🥰
this is so much fun! I just checked out the code, and it's super awesome! Plus, your communication skills are totally amazing 😍
Thanks! Improving day by day.
Hey Dhruv! This is such a clever concept—using a "Ghost Take" mechanism inspired by racing games for public speaking recovery is brilliant, and props for keeping the data locally stored without telemetry bloat.
Quick question out of curiosity: since Electron + React setup handles the 60 FPS rendering and native bindings smoothly for local testing, how would you approach stripping this down if someone wanted to run it as a lightweight, cross-platform CLI/daemon or on lower-end edge hardware without relying on the full Electron desktop shell? Would love to hear your take on keeping that real-time canvas smoothness on a leaner stack!
The Ghost Take idea is the part that makes this more than another score-after-the-fact tool: rehearsing against your own previous pulse curve turns recovery into something you can actually practise, not just read about. I also liked the honesty guard — showing "No measurable pressure moment captured" instead of inventing a flattering metric — and pausing breathing updates while speaking, since motion would otherwise pollute the trace. One question: when the Spacebar-marked moment is a forgotten transition rather than a pricing slide, does the recovery pattern look different enough in the pulse timeline to tell those two kinds of pressure apart?