Rohin Singh - Lead Full Stack Developer, Mumbai, India

Building scalable, high-performance web applications with modern technologies.

Work Experience

Lead Engineer, Northern Trust (Feb 2023 - Present) - Pune, Maharashtra
Senior Consultant, Ernst & Young (Sept 2021 - Feb 2023) - Mumbai, Maharashtra
Software Engineer, BNP Paribas (June 2018 - Sept 2021) - Mumbai, Maharashtra

Technical Skills

Frontend: ReactJS, Redux, TypeScript, JavaScript, Jest, Testing Library
Backend: Java, Spring Boot, Microservices, Kafka, Spring Security, Liquibase
AI / GenAI: GitHub Copilot, LLM & AI APIs, Microsoft Copilot Agents, Prompt Engineering
Testing: Cucumber (BDD), Selenium, JUnit, Mockito
DevOps: Docker, Jenkins, Azure, PCF, GitHub, SonarQube
Databases: Oracle, MS SQL, PostgreSQL
Tools: Postman, Swagger, Jira, Confluence, Agile/Scrum

Projects

Corporate Actions Platform: Operational platform enabling seamless management of corporate actions including Events, Instructions, Cash & Stock movements. Integrated CI/CD pipelines and automation suites.
Compliance Ethics Tool (COMeT): Compliance solution for APAC/UK regulators featuring employee record search, form processing, and dynamic report generation.
Resource Information Management System (RIMS): Full-stack competency tracking system that visualised workforce skills and training gaps through interactive D3.js dashboards.
PropPilot: AI-powered property intelligence platform that aggregates real estate listings, market trends, and neighbourhood data to help buyers, renters, and investors make smarter decisions faster.
MailClarity: Email validation web app that goes beyond syntax checks - performs MX record lookup, SMTP handshake verification, disposable domain detection, and catch-all identification to give a real deliverability verdict.
Dr Sidney Ortho: Professional medical practice website for an orthopaedic surgeon in Mumbai - combines patient acquisition with educational content, multi-hospital affiliation display, integrated social proof, and a direct appointment booking flow.
DrinkFreshFuel: On-demand fresh juice delivery platform for Mumbai - custom blend builder, same-day delivery scheduling, FSSAI-certified quality tracking, and a full e-commerce cart with time-slot booking.
Placemint Group: Corporate web presence for Placemint Group - a clean, conversion-focused marketing site showcasing services, brand identity, and lead capture with a mobile-first responsive experience.

Writing

How we cut release cycles by 40% with BDD and CI/CD

2026-06-18 · 5 min read

On a Tier-0 corporate-actions platform, a Cucumber + Selenium BDD suite wired into CI/CD moved defect detection up 25% and release cycles down 40%. Here is what actually made the difference.

Slow, manual regression testing was our biggest release bottleneck. Every change to a Tier-0 corporate-actions platform meant days of manual verification, and defects still slipped through late in the cycle - the most expensive place to find them.

We started by writing behaviour specifications in plain language with Cucumber, so business analysts, QA and engineers shared one source of truth for what "correct" meant. Each scenario mapped to a Selenium-driven test that exercised the real UI and API paths, not just isolated units.

The step that unlocked the gains was moving those tests into the CI/CD pipeline. Every pull request ran the BDD suite automatically, alongside SonarQube quality gates, so regressions were caught within minutes of being introduced rather than days later in a manual pass.

The results were concrete: defect detection improved by about 25%, and end-to-end release cycles dropped by roughly 40%. Just as importantly, engineers trusted the pipeline enough to ship smaller, more frequent changes - which compounds into better reliability over time.

If you are starting out: keep scenarios readable and business-owned, run them on every commit, and treat a red pipeline as a stop-the-line event. Automation only pays off when the whole team relies on it.

Tags: BDD, Cucumber, Selenium, CI/CD, Quality

A reusable React component system that cut UI defects ~40%

2026-05-09 · 4 min read

Five feature teams were each reinventing modals, drawers and error states. Consolidating them into a small shared component library cut UI bug surface by around 40% and sped up delivery.

When multiple teams build the same UI patterns independently, you get five subtly different modals, five error states and five ways to show a message - each with its own bugs. That inconsistency is both a UX problem and a maintenance tax.

We built a small, well-documented set of shared components - a MessageDrawer, a GenericModal and an ErrorCard - with clear props, sensible defaults and accessibility baked in. The goal was not a giant design system, just the handful of primitives every team actually needed.

Adoption mattered more than the code. We migrated one feature area at a time, paired with the owning teams, and made the shared components the path of least resistance. Once they were easier to use than rolling your own, they spread on their own.

The payoff: UI defect surface dropped by roughly 40%, new screens shipped faster because teams composed instead of rebuilt, and fixes made once in the library propagated everywhere. Consistency became the default rather than something you had to police.

Tags: React, Design Systems, Frontend, Maintainability

Designing resilient Tier-0 microservices with Spring Boot

2026-04-02 · 5 min read

Notes on building business-critical Spring Boot services that stay fast and reliable against Oracle and MS SQL - API design, data-access patterns and the boring things that keep Tier-0 systems up.

For a Tier-0 system, "it works" is not enough - it has to keep working under load, during partial failures, and while it is being changed. On our corporate-actions platform that meant treating reliability and performance as first-class design goals, not afterthoughts.

We kept services focused around clear business capabilities and exposed them through well-versioned REST APIs. Explicit contracts (documented with Swagger) let front-end and downstream teams integrate confidently without reading our internals.

Data access was where most of the latency lived. Optimising queries against Oracle and MS SQL, being deliberate about transaction boundaries, and managing schema changes safely with Liquibase kept response times low and migrations boring - which is exactly what you want.

Operationally, the unglamorous work paid off: consistent logging, health checks, and containerised deploys through CI/CD meant issues were observable and rollbacks were cheap. When something did go wrong, we could see it and recover quickly.

The lesson I keep coming back to: resilience is mostly the sum of small, disciplined decisions - clear boundaries, tested changes, and observable operations - rather than any single clever pattern.

Tags: Spring Boot, Microservices, REST APIs, Architecture

Contact

Email: rohinsingh@outlook.com | LinkedIn | GitHub