Hi, I’m Harsh. 👋
Some call me Bunny, some call me Avi — but most people just call me crazy. I’ll take it.
“Talent hits a target no one else can hit; a genius hits a target no one else can see.” — Arthur Schopenhauer
“Here’s to the crazy ones. The misfits. The rebels. The troublemakers… Because the people who are crazy enough to think they can change the world, are the ones who do.” — Steve Jobs
The Journey So Far
I started my story in Varanasi — a city where, as they say, people come to find eternity. There’s something about living in Kashi that quietly teaches you to think bigger than the ordinary. I graduated from IIT BHU Varanasi in 2018 with a Bachelor’s in Electrical Engineering, and those years remain some of the most formative of my life.
From there, I headed to Carnegie Mellon University to pursue a Master’s in Robotics under Prof. Kris Kitani, diving deep into computer vision, multi-modal localization, and building systems that help machines understand the world the way we do.
But the itch to build something from scratch never really went away.
Building Rephrase AI → Adobe
In 2020, along with Ashray, Nisheeth, and Shivam, I co-founded Rephrase AI — back when text-to-video was more of a research moonshot than a product category. We dove headfirst into synthetic media and generative video, building some of the most compelling AI-driven video experiences the world had seen at the time.
We cloned and created digital avatars of 100+ celebrities — Virat Kohli, Rohit Sharma, Shah Rukh Khan, Hrithik Roshan, and many more. But the moment that truly put us on the map? A Cadbury ad featuring a digitally recreated SRK — which went on to win the Cannes Lions Titanium Award. That one still gives me chills.
Rephrase AI was acquired by Adobe, and that opened an entirely new chapter.
Today — Adobe Firefly & the Future of Video
I currently lead Applied Science and Machine Learning (India) at Adobe as a Senior Manager, working on Adobe Firefly’s text-to-video pipelines for professional and next-generation video creators.
The mission is something I genuinely believe in: making movie creation seamless, and unlocking human imagination at scale. Whether you’re a seasoned filmmaker or someone with a story but no studio — I want the tools we build to get out of your way and let you create.
My research interests have grown with the times. What started as a fascination with Computer Vision and perception has expanded into everything that large foundational models can do — and more importantly, what they could do. I’m in a continuous hunt to explore what’s possible when you combine these models with real-world systems that actually ship.
I think of myself as a generalist at heart (Rephrase taught me that!). I love to code, I love to build, and I care deeply about creating systems that genuinely add value to people’s lives.
Beyond the Code
At my core, I firmly believe — khana, gaana, cinema se bada kuch nahin. Food, music, and cinema. Always.
I made a short film called Stopping by Woods during my undergrad — a project born out of pure love for storytelling. That love hasn’t left me. It just found new expression in building AI that helps others tell their stories.
When I’m not building, I’m reading, writing, talking (a lot), and searching for the philosopher hiding inside the engineer.
Let’s Connect
If you’re working on something that could dent the universe — or even just make someone’s day a little better — I’d love to hear about it.
Feel free to reach out. The best ideas often start with the most unexpected conversations.