Kumail Rizvi
system_metrics
UPTIME_EXPERIENCE
0+YRS
SYSTEMS_DEPLOYED
0LIVE
PRIMARY_RUNTIME
Go
DATA_STORE
PostgreSQL
SECONDARY_STORE
MySQL
NODE_LOCATION
MumbaiIN
About
3 years shipping production Go services. Currently MCA at IIT Patna & IIIT Ranchi.
I write backend systems in Go and PostgreSQL. Four production platforms across payments, real estate, accounting, and society management, each with real money or real property on the line.
My job is making sure the ledger balances, the payment doesn't double-charge, and the listing search doesn't time out at p95. Measurable outcomes, not resume bullets.
I optimize for correctness first, latency second, cleverness never. Connection pools get sized before traffic spikes. Idempotency keys exist on every mutating endpoint. Deploys are boring on purpose.
Based in Mumbai. Open to backend roles at product companies and startups, remote or hybrid.
Futurescape Technologies Pvt. Ltd. · Dec 2024 – Present
Software Developer (Backend)
Mobcast Innovations Pvt. Ltd. · Jun 2024 – Dec 2024
Backend Developer
Unico Connect Pvt. Ltd. · Jan 2024 – Apr 2024
Software Engineer
Destino Infotech Pvt. Ltd. · Jan 2023 – Jan 2024
Developer
Master of Computer Applications
IIT Patna & IIIT Ranchi · 2026 – 2028 (Expected)
Bachelor of Science in Computer Science
University of Mumbai · 2021 – 2024
Systems
4 domains. All live. Each one has a clear blast radius if it goes down.
Domain
Real Estate
Load
100+ events/day
Stack
Go · PostgreSQL · Redis
What breaks if this goes down: Seller onboarding stops. Subscriptions can't renew. Active listings go stale within hours.
Domain
Payments
Load
Transaction processing
Stack
Go · PostgreSQL
What breaks if this goes down: Money stops moving. Every downstream service blocks on payment confirmation.
Domain
Accounting
Load
Financial reporting
Stack
Go · PostgreSQL
What breaks if this goes down: Books close incorrectly. Audit trail breaks. Finance ops reverts to spreadsheets.
Domain
Society Management
Load
Resident ops
Stack
Go · PostgreSQL · MySQL · AWS
What breaks if this goes down: Maintenance requests queue up. Resident communication halts. Billing cycles slip.
Architecture
Most complex system in the portfolio. Idempotent writes, ledger as source of truth, async reconciliation.
Ingress
Ingress
Core service
Supporting system
Idempotency keys and cached duplicate responses.
Source of truth
Ledger and transactions. The only source of truth.
Supporting system
Async delivery with retries and dead-letter fallback.
Scheduled integrity check
Cron-driven verification that matches payment state, ledger entries, and downstream delivery.
Write path
Idempotency key checked in Redis before any DB write. Duplicate requests return cached response.
Source of truth
PostgreSQL ledger table. No balance stored in cache. Every read reconciles against ledger.
Failure mode
Webhook dispatcher retries with exponential backoff. Dead letters go to manual review queue after 5 attempts.
Postmortems
Connection pool exhaustion looks like a slow API until it doesn't. Size pools before traffic, not after an outage.
Idempotency keys aren't optional for payments. One duplicate charge at 2am teaches more than any blog post.
The hardest bugs live at service boundaries — between your code and the database, between your service and the queue.
Engineering Blog
Golang, PostgreSQL, system design. Includes short engineering notes on API design, caching, and deployments.
How to size connection pools, avoid connection exhaustion, and maintain throughput under load.
Implementing clean shutdowns that drain in-flight requests and close resources properly.
Why idempotency keys matter in payment systems and how to implement them correctly.
When to use events, how to handle failures, and patterns for maintaining consistency.
Contact
Backend engineering roles at product companies and startups. Mumbai-based, remote-friendly. Or just say hi if you happen to bump into me.