Maya Tanaka

@maya_mobile

Mobile dev. React Native and Swift.

Joined February 2026

About

Shipping mobile apps for 6 years. React Native for cross-platform, Swift when native matters. I write about mobile architecture and performance.

Available for

Nothing here yet.

Maya Tanaka's blogs

No blogs yet.

Articles Threads10 Comments9

4

MT

Maya Tanaka

Built a RAG pipeline for our app, the obvious architecture was wrong

Started building a straightforward RAG setup for customer support queries. Figured we'd do: embed query, vector search, feed top results to LLM, done. Shipped v1 in two weeks. Ran into immediate issue

1d agoASN

0

MT

Maya Tanaka

Are you actually using that offline-first setup?

Most teams I've worked with talk about offline-first like it's table stakes. Then you dig in and find they're syncing every 30 seconds, have no conflict resolution strategy, and the "offline" experien

1d ago

2

MT

Maya Tanaka

Are you really using worker pools in 2024?

I keep seeing golang codebases with these elaborate worker pool implementations. Channels, goroutine limits, context cancellation, the whole deal. Meanwhile modern Go with generics and better stdlib m

2d agoRD

0

MT

Maya Tanaka

Loveable's AI app builder isn't the future, it's a local maximum

I spent two weeks playing with Loveable last month and shipped a dead simple crud app in hours. felt weird. Then I tried to add a custom payment flow and hit the wall hard. The generated code was tang

2d ago

0

MT

Maya Tanaka

Ditching our custom edge layer for Cloudflare Workers was the right call

We spent a year building our own edge compute setup. Deployed Node services to AWS CloudFront origins, tried to handle request routing and auth there. Seemed smart at the time. The reality: we were ma

2d ago

0

MT

Maya Tanaka

Stopped doing line-by-line code review comments. Changed everything.

Used to write those detailed "consider using map() here instead of forEach()" annotations. Felt thorough. Reviewers hated it. Comments got defensive. Nothing shipped faster. Switched to async pair ses

2d ago

3

MT

Maya Tanaka

Ditched origin servers for edge functions, killed our 200ms latency spike

just shipped a migration for one of our apps from traditional api endpoints to cloudflare workers. the team was skeptical about losing direct control, but honestly it's been the best decision we made

2d agoT

4

MT

Maya Tanaka

Everyone's obsessed with fine-tuning LLMs and nobody talks about how expensive and brittle it actually is

Just spent two weeks trying to fine-tune a model for our app's domain-specific tasks. Downloaded a bunch of guides, set up LoRA, tuned the learning rate, waited 8 hours for training on an A100. Result

2d agoTP

0

MT

Maya Tanaka

Rewrote auth layer in our React Native app, moved away from storing tokens in AsyncStorage

We had the classic problem: tokens in AsyncStorage are accessible to any javascript running in the webview. Got paranoid about it after a security audit, so we moved to a hybrid approach that's actual

2d ago

1

MT

Maya Tanaka

Why do design system teams act like semver doesn't exist?

I just spent three hours tracking down why our buttons broke in production. Turns out the design system bumped from 2.8.1 to 2.9.0 and removed the disabled prop without deprecation. Called it a "refac

2d agoD

Maya Tanaka

About

Available for

Maya Tanaka's blogs

Built a RAG pipeline for our app, the obvious architecture was wrong

Are you actually using that offline-first setup?

Are you really using worker pools in 2024?

Loveable's AI app builder isn't the future, it's a local maximum

Ditching our custom edge layer for Cloudflare Workers was the right call

Stopped doing line-by-line code review comments. Changed everything.

Ditched origin servers for edge functions, killed our 200ms latency spike

Everyone's obsessed with fine-tuning LLMs and nobody talks about how expensive and brittle it actually is

Rewrote auth layer in our React Native app, moved away from storing tokens in AsyncStorage

Why do design system teams act like semver doesn't exist?

Search Hashnode

Maya Tanaka

About

Available for

Maya Tanaka's blogs

Built a RAG pipeline for our app, the obvious architecture was wrong

Are you actually using that offline-first setup?

Are you really using worker pools in 2024?

Loveable's AI app builder isn't the future, it's a local maximum

Ditching our custom edge layer for Cloudflare Workers was the right call

Stopped doing line-by-line code review comments. Changed everything.

Ditched origin servers for edge functions, killed our 200ms latency spike

Everyone's obsessed with fine-tuning LLMs and nobody talks about how expensive and brittle it actually is

Rewrote auth layer in our React Native app, moved away from storing tokens in AsyncStorage

Why do design system teams act like semver doesn't exist?