Portrait of Bala Ramdoss

Bala Ramdoss

Lead Mobile App Architect at Amazon · San Francisco Bay Area

I build AI-powered mobile experiences at Amazon scale. For 10+ years I have worked across Android, iOS, React Native, and web — most recently leading Amazon Lens, a multi-modal visual search experience used by millions of customers, and bringing on-device AI and LLM-powered UX to the shopping app.

Experience

Software Development Engineer III · Amazon Search

2020 — Present

Tech lead for Amazon Lens and More Like This across Android, iOS, and Web.

  • Driving on-device AI integration for Amazon Lens, and architected LLM-powered UX in the mobile shopping app — streaming, skeleton UIs, and progressive loading to make LLM latency feel fast.
  • Designed and launched Amazon Lens, growing its user base 150% YoY through the worldwide rollout of multi-modal search, and launched Lens Live for real-time visual search.
  • Launched 10+ AR features including Virtual Try-On for sunglasses and lipstick, using ARCore, JNI/NDK, and native Android UI; introduced Kotlin to the team.
  • Led the More Like This team across 5 partner teams on web, driving tens of millions in incremental revenue.
  • Bar Raiser for Amazon’s internal A/B testing platform, mentoring experimenters across Search.
  • Hired and mentored 3 interns and supported engineers to 3 promotions.

Software Engineer · Anova Applied Electronics (Electrolux)

2018 — 2020
  • Architected the React Native app, replacing legacy native apps at 99.9% crash-free across 200k+ active users.
  • Built a custom BLE SDK for the Precision Cooker using unidirectional data flow, plus native Android modules for Wi-Fi, NSD, and database migration.
  • Set up CI/CD for automated testing and bi-weekly releases.

Android Developer · Kohl’s Digital

2017 — 2018
  • Built key Android features including Scan & Search; modernized the stack with Kotlin, RxKotlin, Dagger2, Retrofit, and Room.
  • Implemented deferred deep linking and certificate pinning.

Software Engineer · Tesco HSC

2014 — 2015
  • Built mobile app features with Xamarin and Node.js APIs for the app prototype; received the Blue (highest) graduate performance award.

Talks

Sep 2026 · App Growth Summit SF 2026

Building AI Features Users Actually Love

Some AI features become daily habits while others, often more impressive under the hood, get politely ignored. Drawing on real experience building consumer AI features at scale, this session breaks down the traits that earn user love and the quiet ones that push people out of the funnel: what to make fast, what to keep cheap, where friction actually costs you, and what usage patterns reveal about what users truly value — plus a hint at what changes when AI agents start using your app for the user.

▶

Jul 2026 · droidcon USA 2026

How to Lead High-Stakes Initiatives Through Ambiguity

What do you do when you’re handed a high-visibility initiative with a bold but vague vision? This talk covers turning ambiguity into a delivery plan, shipping incrementally while leadership is watching, communicating with directors and the C-suite, and making trade-offs between competing priorities, resource constraints, and organizational bottlenecks.

▶

Jun 2026 · AI Engineer World's Fair 2026

Agent Output Is Not UX: Building the Rendering Layer

A capable agent can find the right information and still leave the user doing the work. This talk is about the layer between the LLM pipeline and the screen: describing the interface as data, constraining component selection before generation, streaming useful component states while the model is still working, and designing for clients you can’t promptly patch — with typed UI contracts and a backend for frontend that turn agent output into native, actionable components.

▶

Nov 2025 · QCon San Francisco 2025 · Innovations in Front-End

Designing Fast, Delightful UX With LLMs for Mobile Frontends

LLMs are slow and unpredictable; mobile users expect instant and smooth. This talk walks through how we close that gap in production: chunking and streaming LLM responses into the UI, managing latency with skeletons and progressive rendering, using prompt engineering to pick the right CX components, and deciding what runs on-device versus on the server — with server-driven UI and BFF patterns holding it all together.

Community

Writing

Contact

Open to conversations about mobile architecture, on-device AI, and speaking. The best way to reach me is email or LinkedIn.

Education

M.S. Computer Science & Engineering
Santa Clara University · 2015 — 2017

B.Tech. Information Technology
Madras Institute of Technology, Anna University · 2010 — 2014