Building consumer and enterprise products that reach hundreds of millions of users.
Engineering leader with 18+ years across web, mobile (iOS / Android), desktop, and backend - with 10+ years leading teams. I currently lead conversational AI and platform engineering at Microsoft Teams, shipping AI-powered experiences (Copilot, Agents, Adaptive Cards) across every client surface. I stay hands-on: architecting streaming AI UX, cross-client schema contracts, and zero-to-one developer platforms.
Impact at a glance
My journey
I started where the hardware met the software. At Cisco in Bangalore, I cut my teeth on embedded video middleware - set-top boxes, real-time streaming, low-latency DVR, and second-screen companion apps. Working across driver teams in the US, Germany, and Canada taught me how large systems actually ship, and I grew from an engineer fixing decoder bugs into a tech lead trusted with quality on a 20M-user product.
Then I learned to build platforms, not just features. At Oracle I took CX Cloud Mobile from zero to ~30M enterprise users - but the work I'm proudest of was invisible to end users: a pluggable architecture, an expression engine, and a heuristic code-generation engine that let entire vertical teams ship ~3x faster without touching the platform internals. That's also where I moved from principal engineer into management, and discovered I could multiply impact through people, not just code.
At Microsoft, I got to do zero-to-one twice. First I built the Teams Developer Platform from scratch - SSO, tabs, bots, message extensions, meeting extensibility - bootstrapping a small team and growing the mobile platform to ~60M monthly active users. Then, as the AI wave hit, I pivoted the same playbook toward conversational AI: shipping Copilot and Agent experiences across web, mobile, and desktop, architecting streaming AI UX and Adaptive Cards at scale, and prototyping the multi-agent patterns that shape where the product is going next.
The throughline: I stay deeply technical while leading at scale, I love the ambiguity of the first version of something, and I keep learning in public - currently an M.S. in Applied AI, three patents filed, and a stream of open-source tools for the developer and AI communities. The next chapter I'm looking for is one where that combination - builder instincts plus leadership reach - matters most.
Experience
- Lead engineering for conversational and AI-powered experiences across Teams' web, iOS, Android, desktop, and backend surfaces.
- Shipped M365 Copilot integration and Agent experiences across all client surfaces, meeting Microsoft Build '23, '24, and '25 announcement milestones.
- Architected Chain-of-Thought API design and streaming AI UX across mobile and desktop - schema contracts, backward compatibility, and parity.
- Own Adaptive Cards streaming, rendering, and schema evolution across all clients (Fluent v9 migration, citations, native optimization).
- Earlier (2019-2022): built the Teams Developer Platform from zero - SSO, tabs, bots, message extensions, meeting extensibility - growing mobile platform to ~60M MAU.
- Built Oracle CX Cloud Mobile from zero to ~30M enterprise users across iOS, Android, React Native, and Web.
- Designed an extensible platform with pluggable modules and a heuristic code-generation engine that let vertical teams ship ~3x faster.
- Built an expression engine and client-side scripting engine for dynamic, customer-extensible UI.
- Architected enterprise SSO & identity (SAML, JWT, client certificates); led a team of 10+ engineers end-to-end.
- Led Cisco Connect (web + mobile, J2EE backend, ~20M users) and Videoscape middleware SDK for second-screen companion devices.
- Designed real-time streaming, recording, and low-latency DVR; built middleware SDKs for embedded platforms.
- Designed authentication / authorization modules (OAuth 2.0, SAML) for video collaboration platforms.
Open-source & side projects
Adaptive Cards MCP
★ 27 - 5 forksMCP server giving AI assistants deterministic tools to generate, validate, and optimize Adaptive Cards for Teams, Outlook, Copilot, and ChatGPT. Lifts LLM card correctness from ~55% to 95%+. Published to npm and listed in the MCP Registry.
PushForge
Swift - macOSZero-config push notification playground for iOS Simulator, Android Emulator, and macOS. 37 bundled templates (13 for AI-agent patterns), real-time APNs byte counter, and payload/platform mismatch detection. ~2,600 lines of native Swift.
AdaptiveCards-Mobile
Swift + KotlinCross-platform Adaptive Cards UI SDK for iOS (SwiftUI) and Android (Jetpack Compose), designed for mobile integration. Actively maintained open-source SDK.
OpenClaw
TypeScript"Any OS, Any Platform" local-first personal AI assistant with a plugin architecture. Companion repos: openclaw-adaptive-cards (native structured-UI plugin) and markdown agent playbooks for engineering-management workflows.
Adaptive Cards IDE & browser tooling
Dev toolsA suite of authoring tools: a VS Code extension with embedded designer and device preview, a Chrome/Edge AI panel for the Adaptive Cards Designer, and a Playwright-based rendering snapshot tool.
voice-trainer
Azure AIOpen-source Teams app for pronunciation training, built on Azure Speech SDK and Azure OpenAI. A practical example of an AI-powered conversational experience.