Kashish Gadhiya
Software Engineer — Full Stack & Backend Engineering
Building Scalable APIs, Cloud Infrastructure & AI-Powered Production Systems
With 1+ year of production experience, I thrive in 0→1 environments — turning complex requirements into deployed, scalable systems that actually hold up. My core expertise sits in backend engineering, REST API design, and cloud infrastructure, with full frontend ownership when the product needs it. I've shipped B2B marketplaces, Stripe payment architectures, LangChain RAG pipelines, and geodata sync tools — all the way from database schema to AWS deployment. My toolkit includes Node.js, Next.js, TypeScript, MongoDB, PostgreSQL, LangChain, Pinecone, and AWS — built to deliver systems that are clean, observable, and production-ready from day one.
What I Bring to the Table
End-to-End Product Ownership
- Sole engineer across 3 production systems at Pardy Panda — owning backend APIs, admin portal, client-facing UI, S3 data pipelines, and AWS deployment without a second engineer.
- Built and shipped a complete AI infrastructure marketplace (InfraCorridors) — backend, frontend, and cloud infrastructure — independently in a single engagement.
- Delivered a full Stripe Connect marketplace from 0→1, covering 4 service types, 10 Stripe APIs, and a complete commission + hold release system.
Payment Systems & Backend Architecture
- Architected a slab-based commission engine across 4 pricing tiers (3%→10%) applied at transfer time with idempotency-key protection — zero double payouts in production.
- Built a 24-hour payment hold system with automated cron-based release every 15 minutes, protecting against no-shows and disputes across 3 service types.
- Integrated 10 Stripe APIs end-to-end — PaymentIntent, Transfer, Refund, Payout, Connect Express Accounts, Account Links, and 4 webhook handlers.
Performance Engineering & AI Pipelines
- Achieved sub-500ms TTFB on data-heavy marketplace pages using React Server Components with Promise.all parallelization — eliminating client-side waterfalls entirely.
- Engineered a TypeScript S3 sync pipeline with MD5/ETag hash diffing — skipping unchanged files on every sync cycle, reducing redundant S3 bandwidth to zero.
- Built LangChain + RAG pipelines with Pinecone for context-aware document retrieval — handling chunking strategy, vector search, and prompt engineering in production.
Key Differentiators
Backend-First, Full-Stack Capable
I think in systems before I think in components — API contracts, database schemas, infrastructure cost, failure modes. The frontend I build on top of that foundation is faster, cleaner, and breaks less. That combination means fewer gaps between layers and faster time to production.
0→1 Builder Under Real Constraints
Every system I've owned has started from nothing and ended in production. No existing codebase to lean on, no senior to escalate to. Just requirements, tradeoffs, and shipped decisions — including cloud deployment, IAM configuration, and Docker containerization handled independently.
AI-Integrated by Default
Not just a developer who dropped in an API call. I've built real retrieval pipelines — designing chunking strategies, configuring vector search with Pinecone, and engineering prompts that return accurate, context-aware results in live production workflows.
Why Work With Me?
I don't need architecture hand-holding. I don't need a separate DevOps engineer to deploy what I build. And I don't slow down when the problem gets ambiguous — I break it into smaller parts and move forward.
I bring startup-level speed with an engineer's obsession over correctness. If you're building something that needs to actually work at scale — a payments system, an AI pipeline, a distributed backend — I've already built versions of all three.
OPEN TO
Collaborating on impactful GenAI / full-stack projects
Speaking about backend system design & RAG performance
Let's connect and build
something
meaningful