AMIT SAINI

Builds backend payment systems and real-time fraud APIs

Punjab, India · Ex Platinum Software · UIET Panjab University

AMIT's brain

Updated 1 hr ago

Drag to spin · scroll to go deeper · click an orb

Page

Payment risk engine
How a 300 ms fraud API achieves recall 0.81 while cutting expected loss
Hospital payments platform
What design patterns enable idempotent money-moving APIs and tamper-evident audit logs
Kafka backend architecture
Why Kafka resilience lets payments survive 60 s outages without failures
Load testing checkout path
Real-time card-payment risk API
What velocity features from Redis Lua scripts power instant risk decisions
Backend engineering for fintech
How backend services move money securely using double-entry ledgers and PCI-style vaults

Growth ML problem
AMIT SAINI @amitsaini1912 · Nov 4
Final-year ML student here, trying to break into AI growth. I built a fraud-risk agent that cut expected loss per transaction from ₹936 to ₹88 by tuning the threshold for cost instead of accuracy. Question for growth folks: what's the ML-for-growth problem you wish someone would solve for you?