KS.Kuldeep Saini
Menu

Kuldeep SainiBackend & Distributed Systems EngineerBuilding reliable systems. Making complex things work.

SDE II at SBS (74Software) · 3+ years

I'm Kuldeep Saini, a Backend & Distributed Systems Engineer working on reactive microservices, distributed data pipelines and performance-sensitive backend systems.

Explore my work Download résumé PDF, 1 page
Currently
Software Development Engineer II
SBS (74Software)
Experience
3+ years · since Aug 2023
Domain
Banking regulatory reporting
Core stack
Java · Spring WebFlux · Apache Spark
Based in
India

Measured results

  • 45+production features shippedsince August 2023
  • 8microservices delivered intoAPIs, Spark jobs, shared libraries
  • 100+regulatory form types automatedon a Spark 4 pipeline I designed
  • -70%bulk export execution time600 s → 180 s, Reactor scheduling fix

About

What I work on, and why it has to be right.

I work on the backend of a banking regulatory reporting platform - the systems that take raw institutional data and turn it into reports a central bank will accept.

Kuldeep Saini, Backend & Distributed Systems Engineer

Most of my time goes into two kinds of software: reactive Spring WebFlux services for form management, bulk exports and authorisation workflows, and Apache Spark jobs that validate and transform millions of financial records. In regulatory reporting a report is either correct and on time, or it is a compliance problem, so reliability is the actual product requirement.

A lot of what I have shipped is about removing failure modes: Spark jobs failing with out-of-memory errors, concurrent writes corrupting records while a background job held them, an export that took ten minutes because of how its tasks were scheduled.

I also own requirements end to end. I work with product managers to turn complex product requirements into scoped Jira tickets with clear acceptance criteria, review code, and mentor five junior engineers.

Reactive services
Spring Boot WebFlux and Project Reactor APIs where thread scheduling decides throughput.
Distributed data processing
Spark 4 batch pipelines over Spark SQL, MongoDB, Apache Iceberg and Trino.
Correctness under concurrency
Locking, leases and job lifecycles across multiple service instances.
Performance and cost
Profiling first, then fixing the cause - and giving the hardware back.

Experience

One company, one promotion, newest work first.

SBS (74Software)

Banking software · regulatory reporting platform · India

Aug 2023 - Present · 3 yrs 4 mos

Aug 2023SDE I2 yrs 5 mos

Jan 2026SDE II11 mos · promoted

Across both roles

  • 45+ production features across 8 microservices
  • 0 batch job failures, after reliability work
  • +50% pipeline throughput
  • 7 → 3 days to deliver to customers

Software Development Engineer II

Jan 2026 - Present

Design ownership of the reporting pipeline, the compliance submission engine and the concurrency model for background jobs.

  • 100+form types

    Designed a pluggable, multi-strategy report pipeline on Apache Spark 4 - step-based stages with interchangeable output strategies. Report generation for 100+ regulatory form types no longer needs manual processing.

  • 0rejections

    Built the compliance report submission engine from scratch in Java, using cell-level filtering rules and structured output generation. No compliance rejections since release.

  • 0corruption bugs

    Implemented distributed record locking in a WebFlux API, tying write access to the background job lifecycle with lease-based expiry. The concurrent data-corruption bugs seen in production stopped.

  • -50%compute cost

    Fixed Spark out-of-memory failures by profiling executor memory, caching reused datasets and moving parsing onto executor threads. Jobs also run 20% faster.

  • 7 → 3days

    Led pre-production upgrades and deployment automation, cutting customer delivery time from 7 days to 3 (about 57%).

  • 5services, 1 sprint

    Delivered a cross-cutting change across five services - APIs, Spark jobs, shared libraries and the frontend - in a single sprint with zero deployment downtime.

Software Development Engineer I

Aug 2023 - Dec 2025

Spark batch jobs, reactive APIs, file ingestion and the shared libraries and CI/CD the other services depend on.

  • 600 → 180seconds

    Cut bulk ZIP export time by 70% in a WebFlux API by correcting task scheduling in the Project Reactor flow. Its Kubernetes allocation dropped from 4 GB / 4 vCPU to 2 GB / 2 vCPU.

  • 4Spark jobs

    Developed four distributed Spark batch jobs processing millions of financial records with Spark SQL, MongoDB and Apache Iceberg.

  • -60%onboarding time

    Built a modular Excel/CSV ingestion pipeline - load, validate, transform, store - reused by three job types without modification. Onboarding a new file format takes 60% less time.

  • 100%issues caught early

    Delivered an institution data import system end to end for banking clients, covering file parsing, multi-layer validation and role-based access. Data quality problems are caught at import instead of surfacing downstream.

  • 8+services

    Extended the shared libraries used by 8+ services with reusable models, error codes and batch step definitions.

  • +30%deploy stability

    Strengthened CI/CD on GitLab CI, Kubernetes and Helm: deployment stability up 30%, production defects down 25%, test coverage above 90%.

  • 5engineers mentored

    Owned discovery for complex requirements with product managers and broke them into Jira tickets with acceptance criteria and dependencies. Reviewed code and mentored five junior engineers.

An index, not a badge wall.

Technical index

  • Set in ink used in the production work described on this page
  • Set in grey also part of my toolkit
Backend
  • Java
  • Spring Boot
  • Spring WebFlux
  • Project Reactor
  • REST APIs
  • Microservices
  • Kafka
  • JUnit
  • Spring Security
  • Spring Cloud
  • Hibernate
  • OpenAPI
  • Python
Data engineering
  • Apache Spark
  • Spark SQL
  • Apache Iceberg
  • MongoDB
  • Trino
  • MonetDB
Cloud & delivery
  • Kubernetes
  • Helm
  • Docker
  • GitLab CI/CD
  • AWS (S3, EC2, EKS)
  • Gradle
  • Git
Fundamentals
  • Concurrency
  • Multithreading
  • System design
  • Low-level design
  • High-level design
  • Data structures & algorithms
  • OOP
  • Agile / Scrum
AI & LLM
  • RAG
  • LangChain
  • Ollama
  • FAISS
  • AI agents
  • Prompt engineering
  • MCP (Model Context Protocol)
Frontend
  • TypeScript
  • Angular

Education & awards

Education

B.Tech, Computer Science Engineering

IMS Engineering College, Ghaziabad

2019 - 2023 · CGPA 8.3 / 10

Awards

  • Kalam Award

    3x

  • Aryabhata Award

    1x

For high-impact contributions to core product development.

Contact

Let's build something meaningful.

If you are hiring for backend, distributed systems or data platform work, email is the quickest way to reach me.