Building products.
Scaling platforms.
Solving hard problems.

Reviews, ratings, Q&A & media · Android · GraphQL · Java · Spark / Kafka / Airflow · GCP

Architecture, Caching & APIs

• Designed and shipped a distributed Memcache cache-aside for item enrichment — seller-reviews p95 about 40% lower; projected 80–95% fewer repeat downstream calls; fails open with negative caching.
• Wrote an 8-option analysis for variant enrichment; recommended staged caching projected to cut downstream load from ~8K to ~400–1.6K QPS without breaking older apps.
• Built OpenAPI-generated service clients and GraphQL resolvers (seller reviews, submit, sampling invite) that form the UGC 2.0 foundation.
• Led item-enrichment investigation across 5+ platform teams; recommendation adopted for the migration path.

Android Performance & Features

• Profiled and rewrote the PDP Reviews component — render time ~40 ms → ~10 ms (~75% faster), janky frames near zero on one of Walmart's highest-traffic screens.
• Earlier variant-reviews caching reduced 1–1.6B API calls/day with ~21% latency improvement and zero business impact.
• Camera image upload and multi-image selection aligned across Android, iOS, and web, with VQA and analytics sign-off.
• Tempo Reviewer Community — architecture plus Android loading, error, and feature-flagged states.
• Closed Android tech-debt detectors and led module/folder structure redesign.

Quality, Tooling & Leadership

• Built PixelMind — in-house visual regression with DINOv2; deployed to Kubernetes with offline model weights from object storage.
• Cut UI-automation false failures by ~30–35%; wrote the Release Validation & Sign-off SOP for Android, iOS, and web; sustained 48+ stable CI runs.
• Built AI-assisted developer tools (automation status skill, feature-flag generator) and team learning guides for the Spark/Kafka/Airflow/Cassandra/BigQuery stack.
• Caught production risks before launch: vendor double-write, V8 heap OOM, canary false rollbacks, plaintext secrets, and expired Kafka certificates.
• On-call, release sign-offs, stress-test ownership; mentor engineers on performance and architecture.

I work with product and leadership through the full lifecycle — discovery, design, delivery, and production risk — not only during implementation.

Examples

• UGC 2.0 dual read/write — designed parallel write/read architecture, caught a vendor double-submission blocker before launch, and drove A/B rollout across 11 surfaces.
• Video reviews — self-initiated from research through production: vendor API deep-dive, progressive upload choice, Android streaming upload, and cross-platform alignment.
• Item enrichment architecture — led investigation with 5+ platform teams; recommended cached item-service path and shipped the OpenAPI/GraphQL foundation.
• Sampling re-engagement — designed and built Spark/Airflow/Kafka reminder pipelines targeting ~33K invitees who never enrolled (87% of invitees).
• AI image search — partnered with data science and search from prototype scoping through mobile delivery planning.

I drive features from first conversation to production — partnering with PM, aligning stakeholders, writing the design, and shipping the code across mobile, API, and data when needed.

Discover & Align

• Partner with PM on epic discovery, feasibility, and sprint planning for video, dual-write, and sampling reminder.
• Coordinate Android, iOS, web, item data, search, notifications, and vendor teams on scope and timelines.

Design & Influence

• Own technical design — dual-write architecture, caching strategy, feature-flag matrices (44 flags documented across 11 surfaces).
• About 10 option-analysis / design docs; found a multi-core platform gap in the Node orchestration framework.

Ship & Grow the Team

• Shipped Dual Read/Write, Video Reviews, and Sampling Reminder from discovery through production readiness.
• Raise the team's bar with PixelMind, release SOPs, data-stack deep-dives, and mentoring.

Rakuten LINK · Rakuten LINK Office · Android · Java / Kotlin · MVVM · SQLite / Room

Rakuten LINK — AI Assistant

• LINK Chatbot module — developed the chatbot module presented as the AI Assistant inside Rakuten LINK, extending the communication app with an assistant-style conversational experience.
• Requirements to implementation — translated product requirements into Android functionality within the LINK application architecture.

Rakuten LINK Office

• Core application development — contributed as a core developer to LINK Office, an enterprise communication product for corporate subscribers.
• Communication capabilities — worked in an app domain covering voice/video calls, group chat, messaging and sharing documents and media.

Android platform work

• Worked with Java/Kotlin, Android SDK/NDK, MVVM, SQLite and Room across communication-focused mobile applications.
• Contributed to offline synchronisation, local data handling and database optimisation for reliable application behaviour.

Product reference: Rakuten LINK Office ↗

WHIDE Customer · WHIDE Partner · Letstrack Mobile · LTRetail · Letstrack AI

WHIDE Customer — Trusted Mobility & Last-Mile Services

WHIDE's customer experience connected people with trusted drivers for scheduled rides and on-demand tasks, with trip visibility and safety features.

• Customer app — developed Android workflows for choosing trusted drivers, arranging rides and requesting parcel, food and shopping delivery.
• Booking and live status — worked on booking details, driver/dispatcher location on a map and pickup or delivery progress.
• Trust and safety — supported a product model centred on trusted driver relationships, trip tracking and shareable ride details.

WHIDE Partner — Driver Operations

• Partner onboarding — built driver-side Android flows for registration, vehicle-document and KYC submission, rate preferences and availability.
• Service fulfilment — supported partner workflows for receiving and completing rides and delivery tasks.
• Two-sided product ownership — contributed across customer and partner applications in a multi-sided last-mile platform.

Letstrack Mobile — GPS Tracking & Messaging

• Real-time GPS tracker — developed Android experiences combining live location tracking for people and vehicles with messaging.
• Location-aware communication — worked on sharing location and media within conversations, connecting social communication with tracking context.
• LTRetail — maintained and developed retailer-facing functionality used to initiate sales and activate tracking devices and service boxes.

Letstrack AI — Google Home Action

• Built from R&D to implementation — researched Google Home Actions and developed the integration from scratch.
• Voice-driven device interaction — enabled users to ask for information about devices and trigger supported vehicle actions, including engine cut-off.
• Backend integration — implemented the conversational integration using Node.js and Dialogflow.

Product references: WHIDE platform ↗ · Letstrack ↗

Digital Concierge · Loyalty & Payments · Node.js Batch Services · Team Coordination

ServiceS Mart — Digital Concierge Platform

• Service marketplace — developed an Android experience for a digital concierge platform where users could request goods and services fulfilled by professional service providers.
• Smart Points loyalty — worked on a loyalty-points concept that allowed users to buy, redeem and transfer points.
• International payment flows — integrated payment gateways for launches in Canada and Mexico, covering the app-side payment journey and validation before release.
• End-to-end ownership — participated in scrum discussions, user-story analysis, sprint planning and estimation; coordinated development and contributed to implementation, design, testing and release.

SMartBatches — Payment & Data Maintenance Jobs

• Node.js batch application — developed and tested backend batch processes supporting payment-system maintenance.
• Scheduled data updates — implemented business-rule-driven jobs to update database information at defined intervals.
• Service integration — used WebStorm and Postman to develop and validate backend workflows associated with payment operations.

Document Management · Live Audio · Customer Support

Docmein — Document Management & Sharing

• Built an Android application for storing photos, files and documents with cloud backup, so users could access their content across locations when connected.

Pundit — Live Audio Commentary

• Developed a live audio broadcasting and listening experience, allowing one user to commentate while others listened from a phone or computer.
• The original project description targeted approximately 1.2-second latency; this is a project target, not a measured production SLA.

User Support App

• Built a mobile support workflow that let users of company applications submit issues directly to support representatives.

Internal Sales · Android Utilities · Commerce · Conversational & Image R&D

• Internal Sales App — developed an internal application supporting company sales operations.
• SmartAlarm — created a reusable Android library to simplify alarm and notification scheduling.
• RoboMart — contributed to an Android storefront for browsing an online catalogue and product information.
• Virtual Chat — built a data-backed conversational experience that returned answers based on user responses.
• Financial Management — developed an application for tracking personal financial outcomes.
• NIR Spectra R&D — explored deriving near-infrared spectral information from images as an image-processing research experiment.

Early Android Product Development · Java · PhoneGap

• Top 25 Contacts — built a PhoneGap app surfacing frequently used contacts and enabling quick contact actions.
• Xtreme Bootcamp — developed an app for viewing class schedules, signing up for classes and checking promotions, challenges, events and studio information.

Independent Publishing · Custom WordPress Theme · Performance · RSS / Editorial Ops · Since 2012 · Rebuild Nov 2026

Founded and maintained a technology publication that has grown into a long-running content property with more than 134,000 published stories. In early November 2026 I audited the live WordPress install and replaced the bottleneck parent theme without deleting the archive or changing permalinks.

• Long-term ownership — built and operated the publication over many years, managing its WordPress site, content workflows, RSS sources, plugins and ongoing maintenance.
• Custom magazine theme — shipped an open-source photo-board theme with desk rails, paged archives, story pages and ad slots that each operator fills from their own account. First-image LCP uses sized images, fetchpriority="high" and a resizer for off-site RSS artwork.
• Measured, not guessed — official mobile PageSpeed Insights after the custom theme: Performance 90, Accessibility 100, SEO 100, LCP 2.7s, CLS 0. Earlier WPVibe lab LCP on the old theme was 13.8s; those two labs are not one continuous chart.
• Constraints kept — posts, permalinks and the AdSense publisher ID stayed. Unused plugins were turned off in the database. Numbers were published only after cache purge on the public URL.

Open sources: technorapper-wp-theme ↗ · repair playbook ↗ · PageSpeed report ↗

Independent Engineering · PHP · WordPress 6.2+ · WP-Cron · OpenAI · Gemini fallback · Instagram Graph · YouTube Data API · X API · GPLv2 · Nov 2026

Designed and shipped a GPLv2 WordPress plugin that turns operator-configured RSS/Atom feeds into source-attributed, AI-reviewed posts, then can optionally distribute a branded graphic, a spoken Instagram Reel and a YouTube Short to the operator's own accounts. It is not tied to one newsroom.

• Fail-closed editorial pipeline — writer and independent reviewer; thin blurbs, deal lists, coupons and recaps are held; policy hits are quarantined; Auto-publish can be switched off for drafts.
• Production operations — database-backed run lock, per-run and daily caps, monthly token/USD budgets, background enqueue for long admin actions, idempotent draft recovery, and a first-party 3-second visible-read counter with signed REST payloads.
• Social distribution — branded stills with rotating layouts, spoken Reels, public YouTube Shorts via Google sign-in (not an API key), and X posting that spends developer credits rather than mixing cost meters. Low-score stills are skipped so the account is not flooded.
• What I learned — treat AI review as a quality gate, not legal clearance; keep credentials off git; separate OpenAI, X credits, Instagram quota, YouTube quota and Fal video spend; retry flagged OpenAI prompts with a shorter excerpt, then Gemini.

Flagship Independent Product · Next.js · TypeScript · Prisma · PostgreSQL · Vercel · AWS Rekognition · Firebase

Designed and built Gatheroly as an independent SaaS product for event organisers, vendors and guests. This product is separate from my professional work at Walmart Global Tech.

• Multi-tenant SaaS architecture — owner, vendor/reseller and client workspaces with tenant-aware routing, isolated event sites, plans, trials and feature entitlements.
• Event management experience — owner dashboards, client workspaces, public event websites, RSVP, guest engagement and event photo experiences.
• AI-powered photo discovery — face-based photo matching with AWS Rekognition, helping guests find their event photos.
• API-first foundation — route, service and repository layers keep business logic reusable and prepare the platform for future mobile clients.

Independent Android Engineering Project · Kotlin · MVVM · Dagger · Room · Coroutines · RxJava · Retrofit · Socket.IO

• Created a reusable Kotlin Android application foundation with a clear MVVM structure and separation of concerns.
• Integrated common application building blocks including dependency injection, local persistence, asynchronous work, networking and real-time communication.
• Structured as a starter codebase to help teams bootstrap Android applications with a consistent architecture.

Android · Java · MVVM · Android Lifecycle · AsyncTask · LruCache

• Built a lightweight Android MVVM boilerplate using platform components instead of third-party architecture libraries.
• Demonstrates asynchronous task handling, lifecycle-aware organisation and Android's LruCache for in-memory data caching.

Proof of Concept · Android · Java · OpenCV

• Explored computer-vision capabilities in an Android application using the OpenCV library.
• Worked with a dedicated face-recognition module and the Android/OpenCV project integration.

Exploratory Project · Kotlin · Compose for Desktop

• Built a Spotify-inspired desktop UI exploration using declarative Compose components.
• Demonstrates a home experience, search and detail views, and reusable desktop UI patterns.

Android Platform Experiment · Kotlin · DEX · Runtime Class Loading

• Explored a proof of concept that downloads a DEX file, stores it in app-private device storage and loads code at runtime.
• Used the experiment to understand Android's runtime loading mechanics; it is presented as a learning PoC, not a production deployment pattern.