Open to backend roles

Bengaluru / --:--:-- IST

Nishanth Gowda

I build reliable financial systems with Go, Kafka & Temporal — and care most about what happens after software meets production.

5+ years shipping to production

Go✦Distributed Systems✦Kafka✦Temporal✦AWS✦MySQL✦Go✦Distributed Systems✦Kafka✦Temporal✦AWS✦MySQL✦

Redis✦Airflow✦Docker✦System Design✦Observability✦Event-driven Architecture✦Redis✦Airflow✦Docker✦System Design✦Observability✦Event-driven Architecture✦

(01)Selected systems

Workthat ships.

9 systems and optimisations from the money-moving core of a lending platform, ordered by impact — each one designed around how it fails, not just how it works.

Event streaming architecture2024

Loan creation, built as a durable event stream.

Designed the core event-driven architecture for loan creation, using Kafka consumers and Temporal Signals to continue financial workflows as asynchronous financier decisions arrive.

GoKafkaTemporalAWS

Events → signals → durable workflow

Async approvals

Replayable failures

Cost optimisation

Loan sync, with over 40% fewer calls.

Optimised the loan sync path to remove redundant work, cutting downstream service calls and infrastructure calls by more than 40% and bringing the cost of every sync down with it.

Go

40%+

Fewer downstream service and infra calls

BeforeBaseline

AfterUnder 60%

Calls per loan sync

Fewer downstream service calls

Lower infra cost

Durable workflows

Multi-tranche handling for LAP loans, on Temporal.

Built the Temporal workflows that handle multi-tranche loans against property, where a single loan is paid out in stages rather than in one disbursal.

GoTemporal

One loan · paid out in stages

Multi-tranche disbursal

Loans against property

Reliability standard2024

A dead-letter queue pattern for every consumer.

Established a team-wide DLQ pattern for Kafka consumers to isolate failures, prevent cascading incidents, and make failed financial transactions straightforward to inspect and reprocess.

GoKafkaRedis

Consumer fails → job scheduler sends 3 retries to DLQ

Zero data loss focus

Team-wide adoption

Collections

A settlement offer calculator, shipped to production.

Built and shipped the calculator that works out settlement offers on outstanding loans.

Go

Outstanding loan in · settlement offer out

Settlement offers

In production

Also shipped.

06 — 09

06Workflow migration2025

Silent batch jobs, turned into observable workflows.

Migrated critical financial batch processes into observable workflows with automated alerts, so failures surface before they become incidents.

15+batch jobs migrated

Airflow / Go / AWS

07Cost optimisation

Redundant Grafana logs, found and removed.

Audited Grafana logs for cost, identified the redundant ones, and removed them.

29%reduction in log volume

Grafana

08Lending journeys

Top-up journeys for off-book financiers.

Built top-up loan journeys for a range of off-book financiers.

Top-up loans · Multiple financiers

Go

09Repayment adjustments2026

EMI and tenure reduction for longer-term loans.

Developed EMI reduction and tenure reduction capabilities for longer-tenure loans, covering the backend workflow changes.

EMI reduction · Tenure reduction

Go / Temporal / MySQL

(02)About — the philosophy

0+

Years building backend systems in production

0+

Critical batch jobs migrated to observable workflows

0

Tolerance for silent failures in financial pipelines

(03)Journey

Where I'veshipped.

I'm a backend engineer focused on reliable financial systems — loan creation, approval pipelines, batch migrations, profiling, and serverless platforms.

Financial workflowsEvent streamsReliabilitySystem design

Read the blog