I build real-time collaboration software, and the teams that ship it.

ProblemMost of Reliance's workforce of around 600,000 collaborated over WhatsApp and phone calls. JioMeet only did video meetings, and features took weeks to reach users.

What I didI'm leading JioMeet's move to chat, audio and video in one product inside JioWorkspace. I merged five repositories (web, Electron, iOS, Android, backend) into one monorepo, moved the team to acceptance-test-driven development with full CI/CD, and put every feature behind a flag for phased rollout. I also set up the team's sprint cadence with product: a prioritised backlog, weekly grooming, sized estimates, spikes for unknowns, demos and retros.

What changedAbout 100,000 people across Reliance now use it. Features ship in 2 days, and with full context for AI tools every engineer delivers end to end across all clients and the backend.

ProblemThe roadmap was weighted toward advanced chat features such as message forwarding and an AI assistant. Usage data told a different story: most users relied on audio and video calls, mostly on Android phones and the desktop app, often on weak networks.

What I didI brought the data to the product managers and argued for putting meetings first: meeting recording, an automatic switch to audio-only when the network drops, a smooth join from mobile, and stability across the board. On weak networks the client now protects what matters most in a work call: audio first, then screen share, with video given up first.

What changedWe reordered the roadmap around the meeting experience and delivered those features first. Many of the recurring complaints about meeting quality on low-bandwidth networks stopped coming in. The chat features followed once the core experience was solid.

ProblemDolby built the conference-room hardware and audio codecs; BlueJeans needed the meeting experience that would run on it and sell through its channel.

What I didI built and led the product with a team of three developers and a QA automation engineer: a meeting app on Dolby's hub, running in a custom Chromium build tuned for Dolby audio, and a remote-control app connected over wired or wireless links for room calendars, joining and in-meeting controls.

What changedIt grew to about 5,000 rooms. CXOs and VPs used it as their dedicated meeting device, so there was no room for failure.

ProblemWhen a user reported a bad call, engineers could only look at server logs. What happened on the user's device was invisible.

What I didI captured WebRTC stats dumps from clients and turned them into Datadog dashboards for per-call debugging, plus trend views and a call-quality index.

What changedEngineering and support could trace a complaint to its cause and watch quality across all calls over time.

ProblemProducts with several apps (customer web, admin portals, desktop, mobile) drift apart when each lives in its own repository.

What I didAs Sequence's first frontend engineer I set up a monorepo for every client app, with shared packages and a design system. At Jio I extended the same pattern to include the backend, and built the JioMeet SDKs on it, which JioMeet itself and external customers use.

What changedShared code stays consistent across apps, and AI tools see the whole system, which is what makes end-to-end AI-assisted development practical for the team.