Hackathons: Kaggle (page 9)
Hackathons · Kaggle: 157 entries in total; this page lists entries 121–135.
GlyphTrack: Detection of Anomalies Under Canopy
Evidence grade: for discovery only; must not by itself support a formal conclusion.
GlyphTrack fuses remote sensing, ML and GPT to pinpoint hidden pre-Columbian geoglyphs beneath Acre’s rainforest canopy.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Amazonia Archaeology Discovery
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Leverage high-res LiDAR and Maya archaeology–trained ML models adapted for Amazonia to uncover hidden ancient structures beneath the canopy.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Archaios- Explore hidden world with Open AI
Evidence grade: for discovery only; must not by itself support a formal conclusion.
An AI-powered exploration platform that bridges the past and the present
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
TerraReveal
Evidence grade: for discovery only; must not by itself support a formal conclusion.
GPT-4.1 + SRTM + Sentinel2 + XGBoost pipeline detects undocumented pre-Columbian earthworks with 90% confidence
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Gemma 2 Swahili
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Fine-tuning Gemma 2 for Swahili language use.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Kyara: Retrieval Augmentation for LLM Fine-Tuning
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Retrieval-augmented data preparation and fine-tuning workflow for a language model.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
ArGemma: Fine-Tuning Gemma for Arabic
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Arabic adaptation workflow for Gemma.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Post-Training Gemma for Italian and beyond
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Post-training workflow for adapting Gemma to Italian and related multilingual use.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Ancient Chinese Expert: Gemma 2>ChatGPT
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Domain and language adaptation of Gemma 2 for Ancient Chinese.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Wendel Big Data Bowl 2025 dataset
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Competition dataset and analysis artifact for pre-snap defensive behavior.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Exposing Coverage Tells in the Presnap
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Methods for identifying pre-snap coverage signals from tracking data.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
TendenciQ
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Modeling defensive tendencies from pre-snap tracking data.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Safety Entropy
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Quantitative evaluation of safety positioning and uncertainty.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Under Cover 2: Predicting Disguised Defenses
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Predicting disguised defensive coverages from pre-snap movement.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24
Dialing Up the Pressure
Evidence grade: for discovery only; must not by itself support a formal conclusion.
Analysis of defensive pressure patterns from player tracking data.
Evidence use: For discovering how winning projects were implemented, their data resources, evaluation methods and application constraints; specific conclusions must be checked against the original project.
Limitations: Hackathon awards are not equivalent to peer review or independent reproduction. Project descriptions may change with the Kaggle page; check code, data and dependency versions before use.
Link last checked: 2026-09-24