QuantaTrade AI — Technology Decisions
Decisions to Make Before Start
1. Blockchain Network
| Option |
Pros |
Cons |
Recommendation |
| Ethereum L1 |
Most liquid, widest ecosystem |
High gas costs, slow |
Only for token listing, not daily ops |
| Base |
Low gas, Coinbase ecosystem, growing adoption |
Newer, less battle-tested |
Recommended for contracts |
| Arbitrum |
Low gas, mature L2, large DeFi ecosystem |
Competitive, many projects |
Good alternative |
| Solana |
Fast, cheap, growing institutional interest |
Different programming model (Rust) |
Only if client requires SPL token |
Recommendation: Deploy $QTRA on Base (EVM-compatible, low gas, Coinbase alignment). Bridge to Ethereum L1 for exchange listings later.
2. Custody Provider
| Option |
Notes |
| BitGo |
Confirmed by client (Richard-HFT, 2026-04-19). Multi-sig / TSS custody, institutional-grade, policy engine, approvals workflow. |
Decision: BitGo. Earlier spec drafts listed Fireblocks — superseded by client's written confirmation on 2026-04-19.
3. KYC Provider
| Option |
Pros |
Cons |
| Sumsub |
Used in QuantaTrade, proven integration, tiered KYC |
Cost per verification |
| Onfido |
Strong enterprise, good coverage |
More expensive |
| Jumio |
AI-powered, fast |
Less flexible |
Recommendation: Sumsub — existing integration patterns from QuantaTrade, proven tiered KYC support.
4. Matching Engine
| Option |
Pros |
Cons |
| Reuse exchange-core2 (Java) |
Already built, proven, fast |
Java dependency |
| Build new in Rust |
Better performance, modern |
Time to build from scratch |
| Build new in Go |
Good performance, simpler |
Less battle-tested for this use case |
Recommendation: Reuse exchange-core2 with modifications. Already handles CLOB, order matching, risk checks. Saves 2-3 weeks of development.
5. Frontend Framework
| Decision |
Choice |
Rationale |
| Framework |
Next.js 15 |
SSR, app router, React ecosystem |
| UI Library |
shadcn/ui + Tailwind |
Modern, customizable, fast |
| Charts |
TradingView Lightweight Charts |
Industry standard for trading |
| State |
Zustand + React Query |
Simple, performant |
| WebSocket |
Native WS + reconnect logic |
Real-time data |
6. Backend Services
| Decision |
Choice |
Rationale |
| API Framework |
NestJS (TypeScript) |
Modular, proven in QuantaTrade |
| Strategy Engine |
Python (FastAPI) |
ML ecosystem, pandas, numpy |
| Matching Engine |
Java (Spring Boot) |
Reuse exchange-core2 |
| Database |
PostgreSQL 16 |
ACID, proven for ledger |
| Time-Series |
TimescaleDB |
Extension of PostgreSQL |
| Cache |
Redis 7 Cluster |
Pub/sub, order books |
| Message Broker |
Redpanda |
Kafka-compatible, simpler ops |
| Search |
Meilisearch |
Symbol search, news |
7. Infrastructure
| Decision |
Choice |
Rationale |
| Cloud |
GCP (or AWS) |
Depends on client preference |
| Container Orchestration |
Kubernetes (GKE) |
Production-grade scaling |
| CI/CD |
GitHub Actions + ArgoCD |
GitOps deployment |
| Monitoring |
Grafana + Prometheus + Loki |
Full observability stack |
| CDN |
Cloudflare |
WebSocket support, DDoS |
| Decision |
Choice |
Rationale |
| Framework |
Hardhat |
Mature, good debugging |
| Language |
Solidity 0.8.x |
Standard for EVM |
| Libraries |
OpenZeppelin 5.x |
Audited, trusted |
| Testing |
Hardhat + Chai + ethers.js |
Standard toolchain |
| Audit |
Certik or Trail of Bits |
Required before mainnet |