SynthGL Hackathon
AI-powered financial data tools built on 44M SEC filing records from 16,647 public companies
01
FinTag
XBRL tag classifier. Enter a financial line item and get the most likely US-GAAP tag. Trained on 1.75M label-tag pairs from 16,647 filers across 68 quarterly SEC datasets.
1,000 tagsfuzzy + model
02
FinAnomaly
Financial ratio anomaly detector. Compare any SEC filer against industry benchmarks using IQR-based z-scores across 10 ratios.
74,993 records7,553 companies
03
Validator
Pre-submission XBRL error catcher. Scans a draft filing for likely mis-tagged line items using AI-powered tag prediction and peer industry benchmarks.
pre-flight checkbaseline + model
Data Pipeline
68 SEC zips->16,647 filers->44M rows->1.75M SFT pairs->QLoRA Qwen3-4B