TrendGen: AI Trend Intelligence Platform
Built for Sila Insights while at StechAI. Social posts are turned into predicted trends, and emerging trends are produced from those posts, current trends, and a large knowledge base, with Prefect ETL pipelines feeding the processing.
Project Overview
Built for Sila Insights while at StechAI. Social posts are turned into predicted trends, and emerging trends are produced from those posts, current trends, and a large knowledge base, with Prefect ETL pipelines feeding the processing.
The Problem
Teams needed to spot emerging consumer trends early from noisy social data, but manually analysing huge post volumes was slow, inconsistent, and impossible to act on in time.
The Solution
A production-shaped ML platform where raw posts flow through embedding, clustering, and LLM naming into ranked trends, while a forecasting layer surfaces emerging signals early, all orchestrated as async jobs behind a real-time React console.
Use Cases
Generation engine
Train, generate, or segment trends with a guided workflow.
Trend explorer
Searchable, filterable trend cards with confidence, keywords, and volume.
Trend deep-dive
Per-trend detail with growth, language mix, and driving posts.
Emerging-signal forecasting
Score early signals before they peak, with evaluation on hit-rate and lead-time.
Async job orchestration
Long-running jobs stream status and progress live to the UI.
Model registry
Versioned topic models keep prediction runs fast and repeatable.
Architecture & Flow
Project Screenshots
What I Delivered
- Unsupervised NLP pipeline with embeddings, clustering, and LLM-assisted trend naming
- Emerging-signal forecasting with vector search and time-series matching
- Async job architecture with live progress for heavy ML runs
- React admin console covering the full trend lifecycle
- Versioned model registry in cloud storage
- Scheduled pipelines and feedback evaluation
Ready to build something useful?
Whether you need a freelance build, a full-time hire, or a short intro call, pick the path that fits and I will respond within 24 hours.