Cactus
Event Ingestion and Growth Analytics Platform
Role
Software Engineer
Duration
2025 – Present
Team
Engineering Team
Status
Active
Overview
Built the event ingestion pipeline and frontend dashboards for a growth platform. The pipeline processes user behavior events in real time, feeds them into analytics views, and powers A/B testing infrastructure. I also integrated third-party marketing APIs for campaign automation.
Problem
The growth team had no unified view of user behavior. Event data was scattered across multiple tools. Running A/B tests required manual data exports and spreadsheet analysis. Marketing integrations (email, ads, CRM) each had separate codebases with no shared interface. The team needed a single platform to track events, run experiments, and manage campaigns.
Approach
- 01Built an event ingestion pipeline that captures user actions, enriches them with session metadata, and writes to PostgreSQL and Redis for real-time and historical queries
- 02Developed React dashboards with D3.js charts showing conversion funnels, retention curves, and experiment results
- 03Implemented A/B testing infrastructure: experiment definition, variant assignment, and statistical significance calculation
- 04Created a unified API integration layer abstracting email (SendGrid), ads (Google Ads), and CRM (HubSpot) behind a consistent interface
- 05Optimized PostgreSQL queries for analytics workloads. Added materialized views for frequently accessed metrics
- 06Worked directly with the growth team to prioritize features based on their experimentation roadmap
Design Decisions
Technology Stack
Frontend
Backend
Data
Integrations
Impact
Event Latency
<1s
Real-time event processing from capture to dashboard display
Query Speed
60x
Dashboard queries dropped from 8-12s to <200ms with materialized views
Integration Speed
1-2 days
Time to add a new third-party marketing integration
A/B Testing
Shipped
Full experimentation infrastructure with statistical significance
Next Case Study
Archv↗