I'm a
I build production AI systems and multi-tenant SaaS that stay fast, observable, and cheap to run at scale.
Full-Stack & AI Systems Engineer
Currently @ Setubridge Technolabs
I care about the unglamorous parts that make software actually work under real traffic — cost, correctness, and reliability at the edges.
I'm a full-stack engineer who lives at the intersection of real-time systems and applied AI. As the sole engineer on a production Shopify AI Shopping Assistant, I've shipped provider-agnostic LLM gateways, bot-to-human handoff over WebSockets, and RAG pipelines that cut cost while raising reliability.
I'm most interested in the unglamorous parts that make software actually work in production — cost accounting down to the microdollar, atomic claims that survive race conditions, and caches that shave milliseconds off every message. I care about systems that are fast, observable, and cheap to run at scale.
Currently pursuing my B.Tech in Computer Science at IIIT Raichur while interning at Setubridge Technolabs. I've solved 200+ DSA problems and I'm always chasing the next hard systems problem.
Full-stack engineer specializing in production AI systems and multi-tenant SaaS. Sole engineer on a live Shopify AI Shopping Assistant.
I design for the whole lifecycle — state machines, race conditions, retries and fallbacks — not just the happy path.
Microdollar-precision cost accounting, dual-layer caching, and token budgets that keep AI features cheap at scale.
Atomic claims, presence dedupe, and reconnect catchup so real merchants never lose a message.
From a Chrome MV3 recorder to an FFmpeg compositor — I ship the whole pipeline, sole-engineer when needed.
From provider-agnostic LLM gateways to real-time WebSocket systems — the stack I reach for, and how deep it runs.
Provider-agnostic gateways, RAG, and agentic pipelines in production.
Fast, accessible React interfaces — from dashboards to embeddable SDKs.
APIs, real-time gateways, and job pipelines built to survive production.
Typed, systems-minded, and comfortable close to the metal.
Relational, document, and vector stores — chosen per workload.
Ship it, run it, keep it up — from container to HTTPS edge.
The everyday kit — plus the platform APIs my products live on.
The patterns that hold multi-tenant AI SaaS together.
Where the résumé bullets became real traffic. A hands-on record of building, owning, and running software that merchants actually depend on.
Sole engineer on a production multi-tenant Shopify AI SaaS, and end-to-end builder of an internal billing platform and an AI marketing-video pipeline.
Shipped a 3-service production Shopify app — Remix backend, a JWT-authenticated Socket.IO gateway, and a React agent dashboard — serving merchants on DigitalOcean with per-shop row-level isolation across 26 Prisma models and 66 REST routes.
Cut LLM cost per conversation ~55% by replacing a monolithic prompt with a 7-state intent engine that loads tier-based prompt modules on demand, with cosine-scored embedding classification and an LRU-cached GPT-4o-mini fallback.
Built end-to-end bot-to-human handoff over WebSockets with atomic SQL-level conversation claims (no double-claim races), shared inbox, multi-tab presence dedupe, and reconnect catchup for messages missed during disconnect windows.
Designed a provider-agnostic LLM gateway over Anthropic Claude and an OpenAI-compatible bus (OpenAI, Groq, Together, Gemini) with unified SSE streaming, tool-schema normalization, and microdollar-precision cost accounting.
Built an internal billing platform (OpsPanel) and an AI demo-video generator from a Chrome MV3 recorder through a GPT-4o scene planner to a Python/FFmpeg compositor.
Production AI systems and multi-tenant SaaS — filtered by domain, searchable by stack. Open any card for the problem, the hard part, and what I learned.
Production multi-tenant Shopify SaaS — sole engineer
Sole Engineer
Internal Shopify pricing, subscription & billing platform
Full-Stack Engineer
Narrated screen walkthrough → auto-voiced marketing video
End-to-End Builder
A real-time window into what I'm building and solving — pulled straight from GitHub & LeetCode.
LeetCode & Competitive Programming
200+ problems solved across LeetCode, HackerRank & GeeksforGeeks.
Not vanity metrics — outcomes shipped in live systems, measured to the microdollar and the millisecond.
Across LeetCode, HackerRank & GeeksforGeeks — a daily habit of hard problems.
Re-architected a monolithic prompt into a 7-state intent engine in production.
Shipped a 3-service multi-tenant Shopify AI app serving real merchants.
Built OpsPanel end-to-end across 16 tables and 15 REST modules.
Where the systems thinking started — a rigorous CS degree paired with the fundamentals that hold real software together.
Have a role, a hard systems problem, or an idea worth shipping? I read every message — the fastest way to reach me is right here.
Prefer your own client? Copy my address or fire off a message — either way it lands in the same inbox.