Yohan (@tyohan) header
Yohan (@tyohan) profile picture

🇮🇩 Indonesian 🦖 Software Engineer 🚲 Weekend graveling | Building AI agents | @inliveapp - audio & video conference 👉🏽 https://teapawt.org/t.co/LFKbahRCMU

Joined July 2008
Lagi nyari AI engineer nih buat dua role ini. Kalau kalian punya background software engineer atau data boleh dicoba apply ya. Key kriterianya sudah sering bikin sesuatu dengan coding assistant, karena nanti bakal dichallenge bikin sesuatu from scratch dengan coding assistant.
16
59
10
605
38,594
I enjoy design events and programs, but this one is the most memorable one for me and the others. Bring web community leaders from different places to connect and learn from each other. youtu.be/8xgspgEyi0Y?si=uITG…
No context Gak bermaksud mancing *when (lagi) ya? yang ada di foto ini dan follower saya, ngacung
1
1
13
2,786
Spec driven might be useful for prompting a new app, but i found out that i don't really use it when working with an existing repo. I mostly prompt it manually. Then i realized most of my prompts are specific enough and work well with small models like Haiku. So use a large model to prompt intention, and use a small model for specific instructions.
2
5
1,005
When working together with a coding agent, we must act like a senior to help a junior, a junior that can work fast, smart enough as long as it has all the context. We must guide them to help them work better
1
2
425
How to work with coding agent effectively. Will post in this thread to recap what i shared at @idswdev . First, understand the principle of coding agent loop and help them to work easier on each step.
2
9
63
2,381
a POC over the weekend
1
385
Submitted
1
3
413
Before & after with Gemini CLI one shot prompt
3
419
Vibe coding this weekend, bikin feature limit broadcast ke specific participants di Inlive Room @inLiveApp . Jadi bisa bikin sesi publik tanpa takut ada broadcast yang tidak diinginkan.
2
1
2
6,129
Dah bisa lihat juga, bisa untuk case-case yang mungkin perlu visual. Jadi bayangin punya asisten AI yang punya mata, mulut, dan telinga dan bisa interaksi langsung tanpa ketik2. Ini logicnya masih perlu diimprove sih.
1
254