The latest articles on DEV Community by Soham Pal (@soham_pal). https://dev.to/soham_pal https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F1598216%2F599be4d4-c637-4dd5-863b-e5b7ce6eb012.png https://dev.to/soham_pal en Soham Pal Sat, 05 Sep 2026 19:48:56 +0000 https://dev.to/soham_pal/i-turned-an-old-60-android-phone-into-a-247-edge-ai-and-s3-server-42w-draw-4g9i https://dev.to/soham_pal/i-turned-an-old-60-android-phone-into-a-247-edge-ai-and-s3-server-42w-draw-4g9i <p>instead of buying another raspberry pi or letting an old xiaomi redmi 9i rot in a drawer, i turned it into a 24/7 edge ai node and local s3/sqlite storage server running on about 1.2w idle and 4.2w peak.</p> <p>hardware:</p> <ul> <li><p>mediatek helio g25 (8x cortex-a53 cores)</p></li> <li><p>4gb lpddr4x ram</p></li> <li><p>emmc 5.1 internal flash + microsd</p></li> <li><p>built-in 5000mah battery acting as a zero-cost ups</p></li> </ul> <p>the software stack:</p> <ul> <li><p>rooted android base running a linux chroot with partial wakelocks so android's aggressive lowmemorykiller daemon and doze mode don't quietly murder background processes.</p></li> <li><p>speech-to-text: whisper.cpp (base.en) compiled with arm neon simd vector instructions on the a53 cores.</p></li> <li><p>llm: llama.cpp running quantized qwen 2.5 (0.5b instruct q4_k_m).</p></li> <li><p>memory governor: an in-process jit governor that automatically evicts idle models back to disk after 75 seconds so the 4gb ram never gets oom-killed.</p></li> <li><p>storage &amp; db: project-isolated sqlite in wal mode plus an s3-compatible blob store with an in-memory hot ram index.</p></li> </ul> <p>the real headache: grid outages &amp; zombie tunnels</p> <p>the power cuts out roughly once a day here. the phone’s battery means the node itself never shuts down, but downstream chaos was the real killer: the router dies, wi-fi drops, and temporary cloudflare quick tunnels would instantly get killed with http 530s or freeze on stale tcp sockets once the connection returned.</p> <p>to make it genuinely self-healing, i moved to a named tunnel with a persistent domain and wrote a background supervisor watchdog. the second wi-fi comes back up post-outage, it flushes dead socket states, tears down zombie listeners, and resyncs the cloudflare edge worker so the database, storage, and inference endpoints recover without me ever touching the phone.</p> <p>actual numbers:</p> <ul> <li><p>qwen 2.5: ~7 tok/sec</p></li> <li><p>local sqlite / s3: sub-millisecond on lan</p></li> <li><p>thermals: sits around 32-34°c with no active cooling <br> <a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fbipvajf9q68y5eevrxza.png" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fbipvajf9q68y5eevrxza.png" alt="Testing Qwen Locally On Edge Server" width="800" height="327"></a></p></li> </ul> selfhosted ai linux programming