// about
Kulshresth Jangid
Senior Backend Engineer · Node.js, TypeScript, Java · Jaipur, India
I design distributed systems, ship production software, and operate what I build. Four years as an engineer across cartech, fintech, and SaaS — working on microservices architectures that handle real traffic, real failure modes, and real consequence when something breaks at 2am. My default orientation is ownership: I define the problem, choose the architecture, write the code, instrument it, and stay on-call for it.
I think in systems before I think in features. Before writing code, I want to understand the consistency requirements, the failure boundary, the scaling inflection point, and the operational cost. I don't optimize for elegance — I optimize for correctness under load and debuggability under incident. Boring, obvious solutions that compose well are better than clever abstractions that hide failure modes.
I founded Kaizex \u2014 a multi-tenant Marketing Operating System \u2014 as a product, not a side project. Four-stage pipeline: Source \u2192 Insight \u2192 Content \u2192 Distribution. Three automation modes. BYO-AI with no model lock-in. Scheduling engine validated at 1M+ tasks/instance. Entry point: tic-tac-toe. Frontend at /kaizex. Backend: smart-server. Designed, deployed, and operated end-to-end.
$ Technical Depth
$ Core Skills
$ Experience
- →Rewrote the Spring Boot data-access layer with targeted query optimization and composite indexing — MySQL p99 latency down 40%, API latency down 35%, user conversion up 12% at the affected flow
- →Designed canary deployment pipeline with automated rollback on Prometheus error budget burn-rate triggers — deployment failure rate down 30%
- →Integrated Pact contract testing into CI — API breaking changes caught at the PR level before staging is ever deployed
- →Led incident response and postmortem process; authored runbooks covering 12 failure modes; on-call MTTR measurably reduced
- →Built JWT-secured partner APIs with Spring Security, enforcing team RBAC at the claim level rather than in application logic
- →Led monolith-to-microservices migration: ran event storming to identify 6 bounded contexts, defined inter-service contracts before any code was extracted, used strangler fig with feature-flag traffic routing for zero-downtime cutover
- →Reduced MTTR by 28% through domain isolation — failures in one bounded context are now structurally contained and cannot cascade across unrelated services
- →Designed distributed saga orchestration for checkout and order workflows with full compensating transaction support and idempotent retry
- →Enforced Pact contract testing across 4 service pairs; breaking changes blocked at CI before reaching staging
- →Optimized MongoDB under high-write load: compound index design, write concern tuning, and read replica routing across all hot query paths
- →Built chatbot automation system using a branching decision engine — increased weekly user engagement by 20%
- →Diagnosed and eliminated N+1 query patterns, redesigned composite indexes, rebuilt Redis caching strategy — p99 DB latency down 40%, API latency down 35%
- →Led LMS migration to zero-downtime using dual-write and progressive cutover — no user-visible impact across the full migration window
- →Moved downstream enrichment calls off the critical response path using CompletableFuture — eliminated synchronous blocking on three high-traffic endpoints
- →Improved Redis cache hit rate from 54% to 87% by replacing per-row TTLs with domain-aligned invalidation and tag-based cache busting
$ Testimonials
Kulshresth was the engineer you'd hand the ugly latency problem to and not have to think about again. The caching redesign on our search APIs cut p99 by 40% — but what stuck with me was that he came with a diagnosis, not just a fix, and the rollout didn't touch our on-call at all.
I've watched Kulshresth go from shipping features to owning entire systems — Kaizex is not a side project, it's a real multi-tenant platform he built pipeline by pipeline. Give him a vague problem and he'll come back with the schema, the failure modes, and the rollback plan.
We didn't need a contractor who just takes tickets — we needed someone who'd push back when a request didn't make sense for the business. Kulshresth did that, shipped fast, and was easy to reach when something broke. That combination is rarer than it should be.
Working across the frontend/backend boundary with Kulshresth was easy in a way that's rare — the API contracts were documented before I asked, and when I found an edge case, the fix showed up the same day instead of a week later.
Kulshresth is the kind of engineer who reads the whole ticket before touching code — asks the annoying questions early so nobody's debugging a wrong assumption three days later.
Calm under a production incident, which is honestly the whole job sometimes. Kulshresth traces a bug methodically instead of guessing, and he writes the postmortem so the next person doesn't have to relearn the same lesson.