Suggestions
Prachie Banthia
Lead Product Manager at rideOS
Prachie Banthia is an experienced product management professional with a notable career in the tech industry. She has held several key positions at prominent companies, showcasing her expertise in product development and leadership.
Career Highlights
rideOS and Gopuff: Prachie served as the Lead Product Manager at rideOS, where she was the first product hire and eventually led the product team.3 Her role at rideOS lasted for approximately 3.5 years, during which time the company was acquired by Gopuff.1 Following the acquisition, she continued her work as the Director of Product Management for the Marketplace at Gopuff.5
AssemblyAI: After her tenure at rideOS/Gopuff, Prachie transitioned to AssemblyAI, where she currently holds the position of Vice President of Product.6 In this role, she has been involved in various initiatives, including leading workshops on user research techniques.4
Skills and Expertise
Prachie is known for her proficiency in:
- Product management
- Identifying growth opportunities in the tech industry
- Leading and developing product teams
- Conducting user research and workshops
Professional Approach
Prachie emphasizes the importance of product managers taking the lead in API specification drafting to ensure alignment within teams.6 She also advocates for individuals who enjoy both people and programming to consider careers in software engineering, solution architecture, or solution engineering.2
As of September 2024, Prachie Banthia's LinkedIn username is prachiebanthia, where she shares insights and updates about her professional journey in the tech industry.
Highlights
I have an innovative design project I need a freelance UX designer to run with. I can’t say much here, but they’d get to think about bringing AI to help companies build better teams.
Reach out if you or someone you know might be a great fit! Would be happy to share more details 1:1 😊
Couldn’t be more proud of AssemblyAI today. We launched Universal-1 – our most capable speech recognition model. Universal-1 is an extraordinary achievement by our research and engineering teams:
-> 13.5% more accurate than models like Whisper and >22% more accurate than APIs from Azure/AWS/Google -> Up to 30% fewer hallucinations than seq2seq models like Whisper -> 71% better speaker count estimation and 14% better word timestamp estimation compared to our prior models -> Just 38 seconds to process 1 hour of audio
My favorite stat is that it has up to 30% fewer hallucinations than seq2seq models. Minimizing hallucinations are critical to the user experience of our customers and the trust between our customers and their end users. A hallucination in a transcript can mean an unusable or inaccurate transcript, bad downstream analytics, and overall limited product experience for endusers. As an AI model provider, bringing forward what’s possible in terms of hallucination rate is critical for us to provide to our customers, so they can build powerful products and better serve their millions of end users.
It’s a memorable time to be developing a product at the forefront of AI. We are learning exactly what matters most to customers when they are building with AI, like hallucination rates. We designed Universal-1 with our customer needs at the forefront, and you’ll see in our research blog post that our model was preferred by humans 72% of the time compared to our previous model.
Every model we put out is more suited to humans, to business use cases, to the real world. Thank you to all our customers that gave us the feedback necessary to build and ship this extraordinary achievement. Excited to share what we come up with next!
