Faisal Fida@faisalfida

PlateSign OCR reads licence plates and road signs from uncooperative images on AMD GPUs. It runs Qwen3-VL-8B-Instruct in BF16 with Hugging Face Transformers on ROCm 10.0, packaged on the mandated rocm/pytorch base image. A resident server loads the model once at container start and warms up the GPU before it reports ready, so each grader call to app.py is a thin standard-library client that sends the image path over a Unix socket and writes the JSON answer. The prompt encodes the transcription rules: keep only the registration number on US plates (drop state names and slogans), keep the province character and letter on Chinese plates, keep every printed word on signs, and join multi-line text top to bottom. Tested in a grader-style run on an AMD Instinct MI300X: 10 of 10 sample images correct across PNG, JPEG and TIFF, 20 s startup, 0.16 to 0.35 s per image, 18.4 GB peak VRAM, 45.7 GiB uncompressed image.

The problem: AI assistants now write a big share of our code, but the request behind each change lives in a chat that gets closed. Git keeps the code and loses the reason: git blame says who committed a line and when, never why it was written or whether it was meant to stay. So the AI's throwaway code (a mock "until the real API lands", a demo script, a feature flag) ships and stays, because nothing reminds anyone to delete it. And nobody can say which AI-written lines a person actually looked at. A story: Sara, a backend developer at a payments startup, finds a mock payment gateway in production. git blame says she committed it three weeks ago, and she has no idea whether it is safe to remove. The solution: Whyline remembers why, at the moment Bob writes the code. 1. When Bob edits a file, small scripts Bob runs automatically save the request and the exact lines. No AI call, no extra cost, about 0.2 seconds. 2. On commit, that record is saved in the git history, attached to the commit. Files are not touched, and the history travels with the code. 3. Temporary code gets a date to go, or goes once nothing uses it. When that happens it is ready to delete, and a pull request check can fail until it is gone. 4. Ask Bob "why is this line here?", "what can I delete?" or "what AI code has nobody changed?". Ask "remove the mock": Bob shows the proof, asks, and deletes only after your yes. Bob cannot edit the history itself. For Sara: Whyline shows her own request ("a mock gateway so checkout works until payments-v2 lands"), that payments-v2 has landed, and that nothing uses the mock any more. Bob removes it after she says yes. Proof, from real runs: In the demo shop, 49 of 96 lines are AI-written and 47 of those were never changed by a person. After payments-v2 merges, the check fails on the mock, and Bob removed it after approval in a recorded run. A public demo repository shows the check passing on one pull request and failing on another. 82 tests, a built-in self check that proves every check can fail, no dependencies, published on npm. Why it is different: Review tools look at a pull request after the fact and guess the intent. Whyline was there when Bob wrote the line and kept the request. And no other AI code tool tracks the temporary code an AI leaves behind until it is safe to delete. Limits: A saved request shows what was asked, not that the code is correct. A person with write access can still edit the history. Changes Bob makes through terminal commands are not recorded. Business model: The CLI is free. Teams pay for Whyline turned on for every developer and a record per release, for example $20 per repository per month

Six Matches. One leaderboard. Build on AMD, earn points and compete for prizes.

🌎 Hybrid Hackathon | 💻 Online Build: 12-17 October 2026 | On-site phase: 17-18 October 2026 To participate in the hackathon, simply click Sign up with AMD. If you are not already a member of the AMD AI Developer Program (ADP), you will need to create an account. The AMD Developer Hackathon: ACT III is a hands-on event for developers, founders, engineers and builders who want to push what’s possible with AI on real infrastructure. 📍 If approved for the on-site experience, you'll also have the opportunity to showcase your project live. Travel and accommodation will not be covered.

💻 Online Hackathon | ⏳ Build: 48 hours | 📅 September 25–27, 2026 The official IBM Bob 2.0 hackathon. Build what's next in AI-assisted development, alongside the developers and experts shaping it. ⭐ Get help from experts throughout the event 🤝 Join a team or build solo 💰 Prize pool: $12,000 🧑🏻‍💻 Register before the Kick-Off Stream to start building with Bob

🚀 Take your chance and build a proactive AI solution. 🌟 Expert mentors will guide you every step of the way. 🤝 Work alone or form a team to build something extraordinary. 💰 $22,250 prize pool for the winners! 🌉 Join us online or in person in San Francisco for an unforgettable experience! 🧑🏻‍💻 Sign up before the Kick-Off Stream to secure your spot!