Senior Software Engineer with 5+ years building Java and Spring Boot platforms for enterprise payroll and compliance at ADP. Architected event-driven, CQRS, and sharded systems on AWS that support 2M+ statements/day for 15M+ employees across 13 countries. Also builds AI tooling that saves 35 dev-hours/week by encoding architectural knowledge.
about
Backend systems at payroll scale
I've spent the last five years at ADP building the platforms behind payroll and tax compliance for millions of employees across thirteen countries — the kind of domain where correctness isn't optional and "eventually consistent" needs a very careful definition. Most of that work sits at the intersection of event-driven architecture, sharded data systems, and keeping a distributed pipeline honest under peak load.
Outside the day job, I write about system design and distributed systems, mentor engineers moving into senior roles, and recently published a book distilling what actually gets asked — and what actually matters — in senior Java interviews.
- Base Hyderabad, India
- Focus Distributed Systems
- Core stack Java · Spring Boot · AWS
- Currently Payroll Intelligence Platform, ADP
- Education MCA, NIT Calicut
ls ./skills
Technical skills
experience
Where I've built
ADP India, Hyderabad — Payroll Intelligence Platform
Oct 2020 – PresentSenior Member Technical (Sr. Software Engineer)
Statement Generation Platform
▸
- Architected a configurable pipeline engine running ~100s of JSON-defined workflows, executing DAG-based data aggregation across dozens of APIs and multiple microservices — generating 2M+ statements/day (peak 7.63M) across 13 countries with zero code changes for new statement types.
- Designed command-range fan-out, partitioning large workloads across pods via SQS-based backpressure — 10x throughput gain through horizontal scaling.
- Built a CQRS architecture: MariaDB writes sharded by country, Elasticsearch reads with a CQEngine in-memory fallback — 1,000+ req/s, 40% faster on simple queries.
- Reached 99.9% pipeline completion via circuit breakers, distributed locking, exponential backoff retry, and per-range fault isolation.
Event-Driven Messaging Framework (Event Machine)
▸
- Designed a Transactional Outbox event platform — events persisted atomically in the same DB transaction as the business mutation, then published asynchronously to SNS/SQS; adopted by multiple teams.
- Built a deduplication state machine eliminating double-processing across pods, with configurable per-listener error strategies, backoff, and Prometheus metrics.
- Designed a Claim Check pattern for large payloads (GZIP → S3, pointer in SQS) and multi-queue horizontal scaling via greedy pod-level queue assignment.
- Led the Kafka → SQS/SNS migration with zero operational overhead, preserving at-least-once delivery guarantees.
Payroll Computation Engine
▸
- Built the gross-to-net payroll engine across 50+ tax jurisdictions, with internal policy execution tools and hot-deployable S3 rule packages.
- Designed a compressed BLOB pattern (GZIP JSON, 10:1 compression) — a single INSERT per calculation instead of 500 normalized rows, with CQEngine in-memory indexing for sub-millisecond reporting queries.
- Architected isolated custom thread pools with OpenTelemetry context propagation, preventing cross-shard interference at peak load.
- Enforced Hexagonal Architecture via ArchUnit — zero framework dependencies in the domain layer, letting the infra layer migrate from JPA to Spring JDBC in isolation.
Payment Orchestration Service
▸
- Designed multi-country payment orchestration — coordinating statement generation, tax filing, money movement, and partner notifications across 13 countries.
- Implemented a bi-temporal data model (transaction time + valid time), supporting retroactive payroll corrections while preserving full audit history.
- Built a choreography-based saga and year-end throttling — queue-based rate limiting with stuck-process detection to prevent downstream cascade failures.
AI Developer Agents & Platform Productivity
▸
- Engineered AI-powered developer agents backed by curated knowledge bases — architecture docs, service interaction maps, failure-mode catalogs, and API contracts — for contextual debugging assistance across microservices.
- Built a structured knowledge-base system encoding tribal knowledge into queryable documents — 35 dev-hours/week saved, new-engineer onboarding cut from 6 to 3 months.
./impact --summary
Selected impact
Education
Leadership
- Mentored 3 engineers — reduced onboarding 6→3 months, established architecture standards across 5+ teams.
- Led production on-call rotations — diagnosed year-end incidents at peak load, implemented capacity planning.
- Authored 20+ architecture decision records; writes technical articles on LinkedIn covering system design and distributed systems.
Certifications
publication
Book
INTERVIEWS
beyond the
boilerplate
questions
A field guide to the Java and Spring Boot interviews that actually happen at the senior level — past the LeetCode round and into JVM internals, distributed-systems tradeoffs, and the follow-up questions that separate a pass from an offer.
Get it on Amazon ↗mail basuki