About
I build systems meant to outlast their original assumptions. This is not a portfolio, but a documented record of architectural decisions, technical philosophy, and the operational realities of maintaining software in production.
My work centers on designing resilient infrastructure where simplicity is a requirement, not a preference.
Engineering Philosophy
Simplicity over cleverness
Complexity is a tax paid over time. Clever code is unmaintainable code — if a system requires a deep mental model to operate, it has failed its primary design requirement.
Documentation as code
Context rots faster than software. Architectural decisions are recorded at the point of origin, because a system without documented constraints is a liability to whoever inherits it.
Optimize for deletion
Every abstraction introduced should have a clear path to removal. Avoid tight coupling to frameworks or patterns that dictate the entire application structure — design for replaceability.
Approach to Building Systems
Understand the business context
Before touching architecture, understand why the system needs to exist and who it fails if it breaks.
Identify the real constraints
Time, team size, and operational maturity shape the solution more than any technology choice.
Design incrementally
Ship the smallest version that proves the architecture, then let production usage inform the next step.
Measure outcomes, not activity
A system is working when it's boring to operate, not when it has the most commits.
Iterate on evidence
Revisit decisions when the data changes, not on a schedule and not out of boredom.
Areas of Interest
Current Focus
Applied AI in Production
Building RAG retrieval and Natural Language to SQL features with LangChain on top of production PostgreSQL systems — with validation to block unsafe or cross-tenant queries.
Security Hardening
Led OWASP hardening for a multi-tenant CRM to a 9.8/10 score — closing XSS, CSRF, and SQL injection gaps across the frontend, API, and database layers.
Production Infrastructure
Diagnosing and stabilizing AWS ECS Fargate deployments — load balancer routing, health checks, and inter-service auth — under live production constraints.
Natural Language to SQL — retrieving CRM data through plain-English queries, with validation to block unsafe or cross-tenant SQL before it reaches production.
Timeline
Software Engineer (AI & Full Stack) — Cloudmate Technologies
Leading OWASP security hardening and building AI-powered features — RAG retrieval and Natural Language to SQL — for multi-tenant CRM platforms, alongside frontend architecture and production deployment ownership.
Full Stack Developer — Investifyd
Owned the entire frontend and roughly 70% of backend for a multi-role Chartered Accountant CRM, end-to-end Razorpay payment integration, a property brokerage CRM, and core workflows for a live fintech investment platform.
Bachelor of Engineering — LJ Institute of Engineering and Technology
Foundational coursework in Ahmedabad, India, before moving into full-stack production engineering.
Beyond Engineering
Placeholder — a brief, honest note on what fills the time away from a keyboard. Replace with real details.