Builder · AI product & engineering

I build AI products.
Then get them into
people’s hands.

I’m an engineer and founder in Bengaluru, working across product decisions, hands-on software and distribution. Currently building to.video, a video workspace where people and agents edit together.

Explore my work Career background

01 / Selected work

Products, and the decisions behind them.

Building now · Private beta

People and agents.
One video workspace.

MakeShorts users wanted more than a generated clip: revise the edit, add music, work with multicam footage. That pushed me toward a broader editing workspace.

In to.video, an agent works beside the team, in the same project. People can edit the timeline, leave feedback, tag the agent to act on it and return to saved versions.

The product decision

Keep the agent’s work editable, reviewable and recoverable.

Explore to.video Watch the product demo
See the working productYouTube
Real product · An editable cut, with the conversation in context.

MakeShorts.ai / Built & grew

From a long recording
to a short worth sharing.

I built MakeShorts.ai to turn long videos into short clips. I also did the early work of finding users: Reddit outreach, AI directories, Google/SEO and direct conversations with creators.

Building a paid product taught me to work across acquisition, customer requests and the editing system underneath. Those conversations eventually helped shape to.video.

Visit MakeShorts.ai

Side projects & experiments

Small things I built to explore audio, voice and agents.

TryClaw

A way to run OpenClaw in familiar chat apps.

Radiofy.ai

An experiment in turning articles and feeds into a radio show.

AudioTools.app

A little browser recorder for narration, retakes and trimming. Original site inactive.

02 / How I work

Product decisions.
Engineering responsibility.

I’ve led engineering and built my own products. My work spans deciding what should be built, designing how it works, building it and finding users.

Start with the work people need to finish.

A feature request is a clue to a bigger job. I use conversations, prototypes and real workflows to decide what to build—and what needs to wait.

Give agents context, tools and a human handoff.

I care about what an agent can see and do, how its output is checked, and how a person takes over. Quality, revision effort, latency and cost all matter.

Build the product and its path to users.

I use agents across development and operations, and OpenClaw for MakeShorts marketing and SEO. I’m exploring how orchestration changes both the way software gets built and the economics of running it.

03 / Writing & point of view

The work should outlive
the agent.

Better models make more work possible. My bet is that lasting product value also comes from the environment around them: shared context, trustworthy actions and work people can keep improving.

That shapes how I’m building to.video’s own agent harness. A cut, its source media, a team’s feedback and the history of decisions have to remain useful when a run stops, a teammate takes over or the model changes.

Make the domain legible

A coding agent reads a repository; a video agent needs picture, sound and an evolving timeline. Context and tools should reflect the actual work.

Keep control in the product

People need to inspect changes, guide the agent and recover earlier work. Permissions and history need to hold as models and runs change.

Count the complete job

I care about the finished result, revision effort, reliability, latency and total cost. That is what connects agent capability to a useful product and a viable business.

I see related design problems across domains: OpenAI’s engineering environments expose software and feedback to agents; Intercom’s Fin brings human decisions into automated support; durable execution preserves progress through failures and waits. The details differ. In video, I’m applying these ideas to shared, editable work—and the judgment needed to make it worth watching.

04 / Background

Experience behind the work.

Head of Engineering · 2021–2022

I built the engineering team and led architecture for Spext’s browser-based audio and video editor. My work covered transcript-based editing, media preprocessing and rendering, alongside product priorities and editing workflows with the founders.

Background
Co-founder & CTO · 2020–2021

artsqft

I co-founded and built a platform for artists to create, manage and sell online performances during COVID-19. The work covered event setup and ticketing, and later extended to film screenings and artist-led workshops.

Visit artsqft
Product Architect · 2018–2019

I rebuilt the architecture of Spext’s browser-based podcast editor, connecting transcript editing with audio processing. The work supported editing long recordings in the browser and was released as Spext 3.0.

Spext 3.0 launch
System Architect · 2015–2018

SureWaves

I led architecture and engineering for Skynet, a programmatic TV advertising platform. It connected TV channel onboarding and ad inventory with media planning for advertisers and agencies, and integrated BARC data for campaign measurement.

SureWaves
Platform Architect · 2014–2015

I worked on the platform architecture of USEReady One, a data governance product.

USEReady
Founder & CTO · 2010–2013

I founded and built Kollabia, an online platform for musicians to collaborate and produce music. The product began as my MTech project at IIIT-B.

Watch Kollabia

EducationMTech, IIIT-B · 2010
Database & Information Systems

SpeakingAWS re:Invent · Las Vegas · 2014
Architecting Reactive Applications on AWS

Outside the editor

I make music, too.

My first vocal song started with college friends in an empty room in 2006: a computer, a mic and instruments I programmed myself. It spread across campus and reached Radio City Lucknow. I’ve since composed and produced songs and the score for the student film Aavrati.

Earlier tracks on SoundCloud
Woh Baarish Mere Khuda Neelkanth – The Sacred Echo Shri Ram (Reprise) Sita Ram Kahiye Sita Ram Kahiye - Longing Peace Film score Aavrati / Teaser

Open original

05 / A conversation

Building an AI product?
Let’s talk.

I’m interested in product and engineering leadership, early product teams and conversations with collaborators or investors around to.video.

Say hello