Welcome Page
AlphaOne's landing page introduces the core value proposition: DeBERTa-ABSA-v2, a model we trained from scratch through 9 iterations to classify per-stock sentiment from Reddit's informal text.
Interactive Inference
Submit any sentence and target tickers for real-time classification by DeBERTa-ABSA-v2. The playground visualizes entity replacement (what the model actually sees) and renders a 12-head averaged attention heatmap for model interpretability.
Operations Dashboard
The terminal view shows a macro sentiment trend chart aggregated across all 88 tracked assets, with daily averages and an operations snapshot of the ingestion pipeline status.
Per-Ticker Sentiment
Drill down into any stock to see its sentiment trend over time, summary statistics, and a sentence-level evidence feed showing exactly which Reddit posts drove each classification.
Platform Architecture
The full data flow across 6 containerized services: React frontend, Spring Boot API, Celery workers, FastAPI inference server, Redis broker, and PostgreSQL, all orchestrated via Docker Compose.
