Slice Bank·Sep 2024 — Present
Engineering Manager
Slice became a bank in late 2024 with no merchant product. Built 4 live payment API product lines across merchant banking, payment APIs, and digital lending.
Key outcomes
- Built Slice's PA/PG merchant business from scratch — 1 customer and <₹5L/month to 4 live product lines (Account Validation, Payouts, PayIns, Mandates) generating ₹70L+/month in 15 months
- PayIn P99 latency: 200s → 7.5s (15×) — rebuilt architecture under live production traffic while simultaneously scaling the product
- AI Oncall Bot: 71% query hit rate, 90s avg resolution vs. 5–10 min baseline — adopted org-wide beyond my squads
- AI OpEx Reporter: 4h/week per squad saved via automated CloudWatch/Sentry → Confluence/Slack pipeline
- Managerial AI Toolkit: 6h/week off EM calendar, 70% retro prep saved across the org
- Cortex — multi-tenant RAG platform shipped at SliceLab hackathon; productionisation underway
The call
The decision to rebuild PayIn was a product and business call — not mine. What I owned was the engineering: running a full architecture overhaul under live production traffic, no downtime, no rollback, while the team was simultaneously building new product lines. That kind of execution — high-stakes, constrained, no clean room — is where I operate best. P99 went from 200s to 7.5s. No major incidents during the migration.