Wafa

• 📝 track your job application 🤖 improve your english speaking (WIP)

[Wafa @mikirinkode]

Halo X 👋
Aku mau share App yg baru aja aku & @jamalakhyer kembangin.

TrackThisJob helps jobseekers (kyk aku hehe😄) to record, manage & monitor job applications directly from mobile phone. App ini jg dilengkapi fitur searching, filtering & quick update.

https://t.co/teQjKBaJk9

[cto.new @ctodotnew]

introducing AI business

a team of agents that work for you

for free

[Gold Tetsola | Creative Intelligence Director @gold_tesla]

codex + opencode go is the strongest coding setup right now.

but anyone who's pushed codex hard knows the wall.

running out of weekly usage on gpt-5.5 mid-build, and your refresh is days away.

it's the best frontier model in the game and it's expensive to keep using on every small task.

i built opencode bridge so you don't have to.

here's what it does:
→ codex spawns deepseek, kimi, and other open source models as native subagents
→ they work the same way gpt subagents do, nothing breaks
→ one paste into codex sets it all up
→ your opencode go key never leaves your machine

here's the wild part:

opencode go is $5 your first month. that gets you the latest open source models.

deepseek v4 pro and kimi k2.6 are beating claude opus 4.7 on coding right now.

the value is insane.

here's how i use it:

native codex (always on):
→ gpt-5.5: orchestrator, final review, critical paths (recovery, auth, schema, migrations)
→ gpt-5.4: bounded implementation and review
→ gpt-5.4-mini: read-heavy exploration, scanning, summaries

through the bridge:
→ deepseek-v4-pro: implementation, debugging, analysis
→ kimi-k2.6: fast repo navigation, review, drafts
→ deepseek-v4-flash: docs, changelog, mechanical edits

gpt-5.5 is the only model allowed on the critical paths. everything else routes to cheaper models.

link in comments.

go build.

[Wafa @mikirinkode]

@adipta_martu my gut feeling is this will be very helpful bro. i was thinking about this job aggregator platform few days ago and then just saw your post. keep it up

[adipta @adipta_martu]

[A Thread]

Jadi, libur lebaran kepikiran buat bikin website aggregator Jobs. Tujuanya cuman biar orang-orang pada bisa jobs searching di satu tempat tanpa perlu pindah2 portal jobs (https://t.co/3Ux4zxqiOX)

1/2

#Tech #Jobs #Vacancy

[Moritz Kremb @moritzkremb]

My OpenClaw "no-AI-slop" video content system

It helps me:
- find ideas
- do research
- plan my content
- write my scripts
- schedule my content
- look at content analytics
- feed it back to self improve over time

...saving me 12h+ every week

Full breakdown & tutorial:

[Wafa @mikirinkode]

@karivarkey thanks for the suggestions! and yeah i have update it

[Wafa @mikirinkode]

https://t.co/fMQArKVOtT I’ve just deployed the app that I developed a few days ago. It’s designed to help me practice my English speaking skills.

There are still lot of bugs, especially when using a browser on a mobile phone 🤣

#BuiltInPublic #BuildWithGemini

[Wafa @mikirinkode]

@kanesiatah96174 THANK YOU!!

[Wafa @mikirinkode]

https://t.co/fMQArKVOtT
I’ve just updated the voice chat app.

- now user able to choose voice female /
- and new AI model Llama3 by Meta that running on Groq

But i still don't know how to fix the bugs when run it on mobile browser 🤔
#BuiltInPublic #BuildWithGemini

[Coolva @coolvatech]

[JOB SEEKERS ALERT]

Siapa yang masih struggle nyatet lowongan kerja yang bejibun itu pake Excel atau manual?

Ga jarang jadi ribet dan malah kelewat kesempatan kan? Been there, done that. Tapi tenang, sekarang ada solusinya!

From this To this

[Wafa @mikirinkode]

@kanesiatah96174 wow well done bro!

[Wafa @mikirinkode]

@isfaaghyth woah never heard about GraphQL before. I’ve just read a little bit about it, and it seems quite different from Retrofit ya Mas Isfa? It looks like don’t need to define each endpoint, just one endpoint using GraphQL query?

[Wafa @mikirinkode]

#MayLearning Day.26

Improve project sebelumnya. belajar text to speech di flutter. Ternyata ada library gratisnya juga ya. Tapi karna gratis hasilnya ya so so

jdi sekarang udh bisa:
- Speech recognition
- Generate response pke Gemini
- Ngomong

[Wafa @mikirinkode]

#MayLearnings Day.25

Nyoba lanjutin project sebelumnya yg speech recognition, sekarang nyoba nambahin Gemini sebagai model LLM untuk ngerespon user.

Dan akhirnya jadi Chatbot simple dengan voice sebagai inputnya.

[Wafa @mikirinkode]

pakai Assistant bisa juga:
- bisa kasih extra knowledge. ex: upload kumpulan pdf jurnal ilmiah. Trus nanti bisa disuruh cari info tertentu berdasarkan file yg ada.
- bisa add custom function
- bisa custom response format jadi json

[Wafa @mikirinkode]

#MayLearnings Day.24

Baru pernah nyoba Assistants di playground nya OpenAI. Biasanya cmn main dibagian Chat + System prompt nya.

Bedanya pke Assistant sama lewat Chat:
- bisa customize persona AI nya, misal mau dia jadi TechRecruiter (cont.)

[Wafa @mikirinkode]

library yang dipake:
- speech_to_text: pastinya untuk rubah speech jadi teks
- highlight_text: untuk highlight text tertentu pada hasil voice
- avatar_glow: untuk efek di mikrofon

credit:
Marcus Ng
https://t.co/VvYpJNEhmH

[Wafa @mikirinkode]

#MayLearnings

Nyoba buat speech recognition app pke flutter. ternyata gk sesulit yg dibayangin karna udh ada library + tutorialnya 😅

padahal semalem nyoba susah susah pke voice recorder trus di passing ke api untuk ubah jadi text to speech. eh ternyata ada librarynya

Day.23

[Wafa @mikirinkode]

logic nya ketika btn ditekan bakal manggil fungsi _listen().

klo _isListening = false, berarti start recording / listening. pantau hasil speech recognitionnya di onResult() dan pakai fungsi .recognizedWords() untuk dpt hasilny

klo _isListening = true, berarti stop recordingnya.

[Wafa @mikirinkode]

#Maylearnings

3 Jam pertama nyoba https://t.co/eeH2JBo5QT

experience yang didapet cukup bagus, mirip vscode jdi cepet adaptasi, untuk build juga gk membebankan laptop hehe

Day.22