Apr 25, 2026
15 min read
On-Device AI Powered Interview – Using Qwen 2.5 1.5B Q4_K_M & Gemma 4
On-device AI on mobile isn't an experiment anymore—it's production-ready. I'm sharing what I learned running a small language model directly on an Android phone using llama.rn. No cloud, no API costs. Just a quantised model living on the device and responding in real time. If you're curious whether this is actually viable for a real app, here's the honest answer.