· 2 min read
StrideDatus — Shipment Cost Optimization Platform
Redesigned a high-latency shipment-routing engine for StrideDatus, cutting computation from 10+ minutes to just seconds through Go-based batching, data-flow redesign, and a TypeScript migration. Improved architecture clarity and frontend stability while reducing infrastructure overhead.
Context
StrideDatus is a logistics analytics startup focused on reducing shipping costs for global freight operators.
Their core engine calculated the cheapest route for every shipment by simulating all possible carrier combinations — but the system took more than 10 minutes per batch and struggled to scale.
Challenges
- Extreme computational complexity — thousands of route permutations per shipment.
- Heavy database I/O and N+1 query patterns.
- A monolithic React frontend built on CRA that was hard to maintain and test.
- Limited backend optimization expertise inside the founding team.
Approach
- Profiled all critical queries, identified redundant data round-trips, and designed a two-phase compute pipeline — batch load → local computation → bulk commit.
- Re-architected backend logic in Go for predictable concurrency and memory safety.
- Migrated the frontend from JavaScript to TypeScript and from Create React App to Vite for faster builds and better modularity.
- Simplified state management using Zustand, improving reliability and test coverage.
Solution
- Replaced incremental database updates with batched writes, reducing I/O by ~90%.
- Introduced typed API contracts between frontend and backend.
- Rebuilt the dashboard as a data-driven visualization layer, making the frontend purely presentational.
- Implemented lightweight telemetry for runtime profiling and performance baselining.
Outcomes
- ⏱ Computation time: 10+ minutes → ~3 seconds per batch.
- 🧩 Architecture clarity: unified backend–frontend contracts and simplified data flow.
- 🚀 Frontend stability: fewer runtime errors, faster builds, improved testability.
- 💡 Strategic impact: founders delayed hiring backend staff and redirected budget to product growth.
