BGMI Post-Match AI Coach -- 2
Budget: ₹150,000 – ₹250,000 INR
I want to automate the way I review my Battlegrounds Mobile India matches. After each game I should be able to feed the raw match recording (or any telemetry you can legally extract) to an AI system that:
• breaks down my individual and squad performance in clear, data-driven terms,
• weaves in strategy suggestions that fit the specific phases of the match—drop, rotations, fights, end-zone choices—and
• closes with practical skill-building tips I can apply next round.
The entire analysis has to come back as a short video summary, not a text report or static chart. Think of a coach pausing the action, overlaying heat-maps, TTK stats, spray patterns, then voicing what I should have done differently.
Key points to consider
– Mobile only: all inputs originate from BGMI on Android/iOS, so any capture, parsing, or model you design has to respect that environment.
– Accuracy over flash: I care more about actionable insights than cinematic edits; if the system says I over-peeked, I need the clip and metric to prove it.
– End-to-end flow: record ➜ upload/stream ➜ AI processing ➜ rendered video back to my device or a shareable link. I don’t want to juggle separate tools.
Deliverables (acceptance criteria)
1. Working prototype that ingests one full BGMI match and outputs a coherent video summary under three minutes.
2. Dashboard or CLI showing the extracted performance metrics that feed the video, so I can validate the numbers.
3. Source code and model weights (or clear instructions if you rely on third-party APIs).
4. Short guide on how to deploy or scale the solution for multiple daily matches.
If you already have experience with computer-vision event detection, mobile game telemetry, or automated video generation pipelines, your expertise will fit perfectly.
• breaks down my individual and squad performance in clear, data-driven terms,
• weaves in strategy suggestions that fit the specific phases of the match—drop, rotations, fights, end-zone choices—and
• closes with practical skill-building tips I can apply next round.
The entire analysis has to come back as a short video summary, not a text report or static chart. Think of a coach pausing the action, overlaying heat-maps, TTK stats, spray patterns, then voicing what I should have done differently.
Key points to consider
– Mobile only: all inputs originate from BGMI on Android/iOS, so any capture, parsing, or model you design has to respect that environment.
– Accuracy over flash: I care more about actionable insights than cinematic edits; if the system says I over-peeked, I need the clip and metric to prove it.
– End-to-end flow: record ➜ upload/stream ➜ AI processing ➜ rendered video back to my device or a shareable link. I don’t want to juggle separate tools.
Deliverables (acceptance criteria)
1. Working prototype that ingests one full BGMI match and outputs a coherent video summary under three minutes.
2. Dashboard or CLI showing the extracted performance metrics that feed the video, so I can validate the numbers.
3. Source code and model weights (or clear instructions if you rely on third-party APIs).
4. Short guide on how to deploy or scale the solution for multiple daily matches.
If you already have experience with computer-vision event detection, mobile game telemetry, or automated video generation pipelines, your expertise will fit perfectly.