BGMI Post-Match AI Coach

I’m building a mobile companion for Battlegrounds Mobile India that chats with players right after a match, breaks down what really happened, and then guides them toward rapid improvement. The flow will be simple for the user—finish a game, open the app, and drop match data or a screenshot; the AI responds in a text chat with a concise analysis, tailored tips, and longer-term performance trends.

What I need engineered is the full pipeline: ingest the match report (API, OCR, or screen scrape—whatever is technically viable), run it through a machine-learning model that recognizes patterns in damage dealt, rotations, loadouts, and engagement timing, and finally generate plain-language feedback plus an auto-updated progress dashboard stored in-app. The chat interface itself can be native or cross-platform (I’m comfortable with Flutter or React Native on the front end), but it must feel instant and lightweight.

Key acceptance points
• The analysis must return within ten seconds on standard 4G.
• Tips should reference the user’s own data, not generic advice.
• Season-long stats, streaks, and heat-map style visualisations should be saved locally and syncable to the cloud.

You’re free to pick the stack—Python with TensorFlow, PyTorch, or a blended approach is fine—as long as model weights and code are delivered, documented, and ready to hand off. If you have prior work parsing PUBG/BGMI logs or building conversational UIs, please mention it; that experience will make integration smoother.

Looking forward to turning great gameplay data into smarter wins. Show More