Product brain. No-code hands. Big builder energy!!
Hey ! 👋
I'm Shivi a Project & Product Manager with 4 years of experience, and honestly, one of the most curious people you'll meet when it comes to new AI products.
I spend a lot of my time exploring what's being built in the AI space, testing tools, breaking things, and figuring out what actually solves real problems vs. what's just hype. No engineering degree, but that's never stopped me from getting hands-on and building.
I believe some of the best product thinking comes from people who obsess over the why and the who- not just the how. And that's where I live.
Excited to connect with builders, founders, and fellow product nerds here. If you're working on something interesting in AI or just want to geek out over the latest launches , let's talk! :)
Replies
Welcome Shivi. I want to hear about the specific AI product you are working on. Would you like to share with me?
I am Yaoshen. I am building in the voice agent product myself. How do you find out the core problem you try to figure out? Thank you.
@lyshen Hey Yaoshen, great to connect!
I’m currently building an AI healthcare automation platform where I’m implementing audio consultation and voice-to-text response functionality to improve patient interaction workflows. I’m also worked on a Smart Apply project focused on AI-driven auto job Apply on LinkdIn matching your Resume ATS score with the Job description.
Usually, I identify the core problem by observing repetitive manual tasks, user frustration points, and workflow gaps. I spend time understanding what users are already struggling to solve manually, that’s often where the real opportunity exists.
Would love to hear more about your voice agent product too!
@shivi_sharma6Â It reminds me a TV series named "The Pitt". It shows that the voice interaction is perfect for patients. However, transcription error and typos are definitely challenge for medical records. But using AI for summaries rather than precise diagnostics sounds very practical. Is your product focusing on these post-consultation summary scenarios?
@lyshen That’s a great reference, The Pitt captures the healthcare environment quite realistically! And I completely agree, transcription accuracy and medical context are major challenges, especially when records need precision.
My current focus is less on diagnostics and more on improving the consultation and post-interaction workflow. So yes, AI-generated summaries, voice-to-text capture, and making patient-provider communication more seamless are definitely part of the direction. So I see AI more as an assistive layer that reduces friction rather than replacing medical judgment.
Curious in your voice agent work, how are you approaching accuracy and trust in conversations?
@shivi_sharma02Â It is not an easy job to approach accuracy and trust in conversations. I have a similar work, which I haven't launched on PH yet. In this project, I make a dataset in which first I extract the facts from each sentence, so it can calculate the accuracy and recall metrics. Not easy, facts can have different expressions, the implementation of score needs skills. Then, the third layer is how to organize the facts. I focus on the metrics of the topic clusterings. I don't label too much data because it is not an easy job. What about you?