The owl mascot waving hello with one wing

Riken PatelSenior AI & Mobile Engineer

I fix, ship, and scale React Native + AI products that can’t afford to break.

Ten years of production apps at Spin (Ford Mobility), Intuitive Surgical, First Tech FCU and Blue Cross Blue Shield of Michigan. Now building Intervoo and Jishu Labs, and taking on AI and mobile work for founders and teams.

  • Available for new projects
  • Fresno, California · Pacific Time
The owl mascot waving hello with one wing
riken.config.ts
  1. export const riken = {
  2. role: "Senior AI & Mobile Engineer",
  3. based: "Fresno, California",
  4. mobile: ["React Native", "Swift", "Kotlin"],
  5. ai: ["Streaming LLMs", "Voice", "RAG"],
  6. web: ["Next.js", "TypeScript", "Supabase"],
  7. building: ["Intervoo", "Jishu Labs"],
  8. available: true,
  9. }
React NativeStreaming LLMsReal-time voiceSwiftKotlinNext.jsRAGTypeScriptSupabaseNative modules
Spin (Ford Mobility)Intuitive SurgicalFirst Tech FCUBlue Cross Blue Shield of MichiganIntervooJishu Labs

What I do

Fix it. Ship it. Scale it.

Clients usually hire me for one of these three. Pick the one that sounds like your week.

When it’s breaking

Fix.

Crashes, ANRs, memory leaks, state bugs, 60fps regressions and CI breaks. I triage, find the root cause and ship the fix.

  • React Native crash and performance triage on iOS and Android
  • Performance audits: cold start, list scroll, API p95, bundle size and memory, with a written report of ordered wins
  • App Store launch rescue: review rejections, missing entitlements, crashes on submission

When it needs to exist

Ship.

MVPs from zero to beta in 4 to 8 weeks, in React Native, Next.js and Supabase or your stack. Weekly demos and a fixed scope.

  • AI features in production: streaming LLMs, real-time voice, RAG and agents
  • iOS, Android and web from one engineer who has shipped all three
  • Auth, billing and data on Supabase, Postgres Row Level Security and Stripe

When it has to hold up

Scale.

Architecture for AI and mobile products with real latency budgets: streaming, offline-first sync and native modules in Swift and Kotlin.

  • Multi-tenant Postgres with Row Level Security, enforced at the database
  • HIPAA, PCI, biometrics, SSL pinning and audit trails, shipped at Intuitive Surgical, BCBSM and First Tech
  • Senior consulting: architecture reviews, tech due diligence and hiring bars

Selected work

Things I built and shipped.

Two products I run, four AI Lab apps you can open right now, and three years of React Native at Ford’s micromobility company.

AI productFounder2024 to now

Intervoo

Live AI interview copilot

It listens to a live interview, recognizes the question and streams a personalized answer grounded in the candidate’s own stories. Audio is processed in the browser and never stored.

  • Next.js
  • Streaming LLMs
  • Real-time voice
  • Supabase + RLS
  • Stripe

Founder & Principal Engineer · solo build

Dev tools2024 to now

Jishu Labs

AI-augmented developer tools suite

Eight AI developer tools on one shared backend, one design system and one sign-in. New tools take days instead of weeks because the foundation is already solved.

  • SenseiAI Code Review & Comprehension
  • KeisanAI Project Estimation
  • KiokuDecision & Context Memory
  • KikuVoice & Audio Intelligence
  • KozoAI Database Schema Generator
  • ShivoAI Spec & PRD Generator
  • MichiLearning Paths & Roadmaps
  • JumbiAI Readiness Assessment

Principal AI Engineer (contract)

AI Lab

Four live apps. Open one.

Small, sharp builds that each run in your browser at their own address, with a full case study behind every one.

Conductor's owl, waving a baton at three small robot agents
Conductor console in pen and ink: a finished run engraved as an orchestral score, with the planner, three workers and the critic as staves and a verified stamp

Conductor

Multi-agent orchestration engine

A planner splits the goal, tool-using workers run in parallel and a critic checks the answer against its sources before anything ships.

Librarian's owl holding a stack of books and a glowing open book
Librarian console in pen and ink: a question filed on an index card, its score sheet, and the catalogue cards pulled from the drawers with the cited one marked

Librarian

RAG knowledge assistant

Answers questions over your own docs with hybrid search and reranking, cites the exact source for every claim, and refuses when the context is thin.

Converse's owl wearing a headset next to a sound-wave speech bubble
Converse console in pen and ink: a live call drawn as a comic strip, with speech balloons, a tool call inset and the inked call waveform

Converse

Realtime streaming voice agent

Streaming speech-to-text into a tool-using model into streaming speech out, with barge-in, built for a natural back and forth.

Verdict's owl in a judge's collar with a gavel and golden scales
Verdict console in pen and ink: the golden set as a court docket with PASS and FAIL stamps, the judge owl and a blocked release-gate stamp

Verdict

LLM eval & release gate

Golden datasets, LLM-as-judge scoring and regression gates in CI, with cost, latency and quality tracked on every run.

MobileFord MobilityFeb 2021 to Feb 2024

Spin

The rider app for Ford’s micromobility company

Three years on the production iOS and Android apps: React Native features across the full ride, native modules in Swift and Kotlin for location and Bluetooth, and the CI that got it all to the stores.

  • Live map with clustered scooter markers and predictive tile caching
  • Bluetooth scooter unlock with an offline-safe retry
  • Deep links across sign-in, the trip lifecycle and transit partner apps
  • React Native
  • Swift
  • Kotlin
  • Google Maps SDK
  • BLE
  • Bitrise

Senior Software Engineer · San Francisco

Experience

Ten years in production.

From hybrid healthcare apps in 2016 to real-time AI products today, across mobility, surgery, banking and insurance.

  1. 2024 to nowNow

    Intervoo

    Founder & Principal EngineerRemote

    A real-time AI interview copilot built end to end: streaming LLM pipelines, live voice and a privacy-first design with zero audio storage.

    • LLM streaming
    • Deepgram
    • Next.js
    • Supabase
    • Stripe
    Case study
  2. 2024 to nowNow

    Jishu Labs

    Principal AI Engineer (Contract)Remote

    A multi-product AI platform with a shared backend, design system and auth layer, and reusable LLM and voice patterns across independently deployed apps.

    • LLM systems
    • RAG
    • TypeScript
    • Postgres RLS
    • Netlify
    Case study
  3. 2021 to 2024Feb 2021 to Feb 2024

    Spin (Ford Mobility)

    Senior Software EngineerSan Francisco, CA

    Rider-facing features for the production iOS and Android apps in React Native, with native modules in Swift and Kotlin and CI on Bitrise, Azure DevOps and GitHub Actions.

    • React Native
    • Swift
    • Kotlin
    • Google Maps API
    • XCUITest
    Case study
  4. 2018 to 2021Nov 2018 to Feb 2021

    Intuitive Surgical

    Mobile Application DeveloperSunnyvale, CA

    Secure mobile and web apps supporting surgical workflows in the da Vinci robotic surgery ecosystem, with offline-first data flows and OKTA sign-in.

    • React Native
    • Ionic
    • OKTA
    • AWS Lambda
    • Node.js
    Case study
  5. 2017 to 2018Jul 2017 to Oct 2018

    First Tech Federal Credit Union

    Mobile Application Developer (React Native)Mountain View, CA

    Mobile banking features: payments, transfers and peer-to-peer payments with Zelle, on a Redux architecture with Jest coverage.

    • React Native
    • Redux
    • Google Maps API
    • Jest
    Case study
  6. 2016 to 2017Aug 2016 to Jun 2017

    Blue Cross Blue Shield of Michigan

    Web Developer (Ionic, IBM MobileFirst)Detroit, MI

    Hybrid mobile apps for healthcare services on IBM MobileFirst, with biometric sign-in through Cordova.

    • Ionic
    • AngularJS
    • Cordova
    • IBM MobileFirst
    Case study
  • M.S. Computer Science, Northwestern Polytechnic University
  • B.Tech Computer Science, Gujarat University

How I work

From first call to shipped, in four steps.

No vague timelines, no scope creep, no surprise invoices. A clear path from the first call to a product your users rely on.

  1. Discovery call

    30 min, free

    We talk about your goal, your users and what success looks like. No pitch, just a clear answer on whether this is a fit.

  2. Scoped proposal

    within 24 hours

    A written proposal with a fixed price, weekly milestones and exact deliverables. No hourly estimate that balloons.

  3. Weekly demos

    every Friday

    A working increment every Friday. You approve the direction, and nothing reaches production without your sign-off.

  4. Launch and support

    30 days included

    I ship it, hand over clean code and docs, and stay on for 30 days to fix anything production turns up.

Four rules I’d stake a project on.

The gap between a demo and a product is where most AI ideas die.

The useful engineering isn’t the first 80 percent. It’s the last 20, where latency, cost, eval coverage and failure modes decide whether the thing survives a Tuesday.

Constraints early are a gift. Flexibility early is a trap.

Privacy requirements, a strict latency budget, a fixed scope. The projects that defined those on day one shipped faster and cleaner than the ones where everything was negotiable.

Premature abstraction costs more than writing the code twice.

I’d rather ship the obvious version, see what actually matters and factor later. The second version is always sharper than the one written in the dark.

Enforce hard rules at the deepest layer they apply.

Tenant isolation belongs in the database, not the API. Accessibility belongs in the component primitive, not QA. One layer of real enforcement beats three layers of polite reminders.

Questions

What clients ask first.

Buying-stage questions from founders, CTOs and heads of engineering, with straight answers.

Ask your own
Can you work with our existing team?

Yes, probably more often than solo. I slot in as a senior IC or tech lead on your existing React Native, AI or full-stack team, pair with your engineers in code review and planning, and leave the codebase in a state your team can keep shipping without me. I default to your tooling, your coding style and your PR process, not mine.

Can you audit before rebuilding?

Almost always the right first step. A one to two week paid audit (cold start, p95 latency, crash and ANR rates, bundle, memory, CI, architecture) usually finds weeks to months of wins that don’t need a rewrite. You get a written report with ordered fixes, effort estimates and a clear go or no-go recommendation before you commit to anything larger.

Can you work under NDA?

Yes. Mutual NDA on day one: happy to sign yours, or send you mine. I’ve shipped under HIPAA (Intuitive Surgical, BCBSM), PCI (First Tech) and enterprise IP NDAs. That means I know audit trails, encrypted storage, SSL pinning rotation playbooks, and when a requirement is security theater versus actual risk.

How fast can you start?

Usually within one to two weeks of a signed SOW, sometimes the same week if a current engagement is wrapping. An audit or emergency triage (launch rescue, crash spike, App Store review block) can typically start within 48 to 72 hours.

Do you handle App Store release?

Yes, end to end. Apple and Google Play submissions, provisioning, TestFlight and internal testing tracks, policy and privacy declarations, entitlements, review responses for rejections, phased rollouts and post-launch crash monitoring. I’ve walked apps through the full release cycle at Spin, First Tech and BCBSM, including the awkward rejection-and-appeal paths.

Do you work with startups and enterprise?

Both, and they’re very different gigs. Startups: fixed-scope MVPs in weeks, pragmatic architecture, founder as PM. Enterprise: work inside your process, with security reviews, SSO, change management, longer rollouts and stakeholder alignment. Ten years across both sides (Ford, Intuitive Surgical, BCBSM and First Tech on the enterprise end; Intervoo and Jishu Labs on the startup end) means I can switch between the two depending on the week.

Contact

Have a real problem? Let’s scope it in 30 minutes.

An AI feature to ship, a mobile app to rebuild, or a second opinion on your architecture. Free 30-minute call, no pitch, clear next steps either way.

Email
Based in
Fresno, California · Pacific Time