Bracket Report × Sertechs
From Static Charts to AI-Powered Financial Intelligence
Client
Industry
Services
Key Result
Our Portfolio
The Challenge: Modernizing a Rudimentary System
When Bracket Report approached Sertechs, they were using a rudimentary reporting system: Python scripts, Matplotlib-generated SVGs, and cron-based HTML page generation. The platform lacked a backend, user interactivity, and scalability. Our mission: Rebuild the reporting engine into a modern, intelligent financial platform from backend infrastructure to client-facing dashboards.
The Evolution of Bracket Report
Then: The Old Ways
- Static HTML via Cron Jobs
- Matplotlib Charts as SVGs
- No Authentication or Backend
- Manual CSV to BigQuery
- Manual Symbol Matching & Peers
- Fixed Reports with No Interaction
Now: Powered by Sertechs
- Dynamic Django Web App on Google App Engine
- Interactive JS Charts via Keen Theme & Custom JS
- Firebase Auth for Clients, Django Admin for Ops
- Automated Pipelines via GCS + SFTP + Cron Schedulers
- AI-Powered Symbol Parsing + Peer Auto-Suggestions
- Dynamic Filters, Real-Time Views, AI-Driven Client Insights
What Sertechs Built
Backend Architecture
Built using Django and hosted on Google Cloud App Engine. Uses Firestore (for client profiles) and BigQuery (for financial data). Automated pipelines using Google Cloud Storage (GCS) + SFTP feeds.
- Django on Google App Engine
- Firestore for client profiles
- BigQuery for financial data
- Automated GCS + SFTP pipelines
Admin Dashboard
Empowering operations with efficient management tools.
- Client Onboarding: Auto-validates exchanges/symbols, integrates symbol look-up API.
- Peer Management: Manage peers with built-in symbol parsing and BigQuery sync.
- News Management: Review/categorize news from MBOUM API and push to news_ca.
- Data Fetch Triggers: Admins can pull fresh data on-demand per client.
Client Dashboard
Delivering actionable insights to end-users.
- Authenticated using Firebase.
- Reports load dynamically based on each client’s assigned symbol/exch.
- Frontend powered by Keen Theme + custom JS layer.
Agentic AI (In Progress)
We’re developing a smart assistant for Bracket clients:
- Users will ask questions in natural language (e.g., “How did short sales perform last week?”)
- AI will query Firestore + BigQuery and return: Text answers, Supporting charts, Trend explanations
