Open to backend roles

Sunny Kr Singh.

Java / Spring Boot Backend Engineer

I build REST APIs, security and data models that hold up under real use — currently working on ResolveAI.

Bengaluru, India

View WorkResume
Email Sunny ↘

System in
focus.

resolveai / triage-pipeline
Open case study ↗
  1. POST /tickets
  2. Outbox
  3. Worker
  4. PII Redaction
  5. LLM
  6. Policy
06 /

Deterministic rules compute priority and routing.

About

I'm an MCA student and a backend-focused developer. I build real systems with Java and Spring Boot, and I care about the parts that have to be right: REST API design, authentication and authorization, database modelling, and how the pieces fit together.

Most of what I know comes from building projects end to end: a delivery backend with a state machine, a payments API with verified webhooks, and now ResolveAI, where I'm exploring how AI fits into a backend without taking over its decisions.

Experience

Where I've built.

Redalis

Jun — Sep 2026
Web Development InternBangalore, Karnataka

I owned the backend of a content and website-management platform: the authentication system, the data model behind its multi-level organizational hierarchy, and the caching and rate limiting that kept repeated reads and abusive requests off the database.

6+ backend modulesJWT auth with refresh tokens and RBACHierarchical PostgreSQL data modelRedis caching + rate limiting

The full task-by-task breakdown lives in the résumé.

Open résumé ↘

Selected work

Case
Studies.

3 Systems
Full write-ups inside

01AI-Assisted Backend

ResolveAI

AI-Assisted Helpdesk & Incident Triage Platform

A helpdesk backend where the model produces triage signals and deterministic policy turns them into priority and routing. The AI suggests; the system decides — so every decision stays explainable and testable. In active development; nothing here is presented as deployed.

Java 21 • Spring Boot 3.x • PostgreSQL 16In Development
→
02Delivery Systems

Hyperlocal

Multi-Tenant Delivery Management Backend

A multi-tenant delivery backend: customers create shipments, the API assigns the least-loaded available agent, and every status change moves through a state machine into an audit trail. A React frontend covers live tracking and agent dashboards.

Java 17 • Spring Boot • Spring SecurityCompleted
→
03Payments

E-Commerce REST API

Catalog, Checkout & Payments Backend

A production-style commerce backend: catalog, cart and orders with Stripe Checkout. Order state only changes after a webhook's signature is verified.

Java 17 • Spring Boot • Spring SecurityCompleted
→

Engineering decisions

The reasoning behind the systems.

01
Why PostgreSQL?
SKIP LOCKED, partial unique indexes, JSONB, full-text search and pgvector in one engine.
02
Why Redis?
Idempotency keys, rate limiting, caching and short-lived locks.
03
Why an outbox?
Ticket persistence and event creation must be atomic.
04
Why a modular monolith?
There's no independent scaling or deployment requirement yet.
05
Why pgvector?
Vectors stay transactionally close to application data.
06
Why deterministic policy after AI?
AI provides signals; business decisions stay explainable and testable.
07
Why state machines?
Invalid transitions become impossible instead of merely discouraged.
08
Why refresh tokens?
Short-lived access tokens limit exposure without forcing frequent logins.

Stack

Grouped by role, not ranked.

Java

  • Spring BootPrimary framework across every project.
  • Spring SecurityJWT authentication, RBAC and method-level authorization.
  • JPA / HibernateEntity modelling on MySQL and PostgreSQL.
  • REST APIs43 endpoints in Hyperlocal; 6+ modules at Redalis.

Data

  • PostgreSQLHierarchical models, SKIP LOCKED, full-text search.
  • MySQLHyperlocal and E-Commerce persistence.
  • RedisCaching, rate limiting, idempotency.
  • pgvectorHybrid retrieval in ResolveAI.

Engineering

  • JWT / RBACAccess + refresh tokens with role checks.
  • State MachinesDelivery and ticket lifecycles.
  • WebhooksStripe signature verification.
  • Async ProcessingOutbox + worker pipelines.

Tools

  • MavenBuild tooling.
  • FlywayVersioned schema migrations.
  • DockerLocal infrastructure.
  • TestcontainersIntegration tests against real databases.
  • OpenAPI / SwaggerAPI documentation.
  • Git / GitHubVersion control.

AI Systems

  • Spring AILLM integration from Java.
  • RAGRetrieval-backed, citation-enforced drafts.
  • Vector Searchpgvector + full-text with RRF.
  • AI EvaluationMeasuring draft quality and coverage.

How I work

01

Build for correctness

State machines, validation, transactions and constraints should prevent invalid states.

02

Keep critical decisions deterministic

AI can help with unstructured input; important business decisions stay explainable and testable.

03

Security by design

Authentication, authorization, tenant isolation and input validation are part of the architecture.

04

Measure what matters

Latency, queue depth, SLA performance, cache hit rates and AI costs should be observable.

Academic record

MCA

Jain (Deemed-to-be University), Bengaluru

SGPA 8.739 — 2025 — 2027

Foundation

BCA

St. Xavier's College of Management & Technology, Patna

CGPA 8.29 — 2022 — 2025

Want the complete overview?

Download Resume