Hi, my name is
Sanjay
I build AI-powered backend systems.
I'm a Backend & AI Engineer and Founding Engineer at Comergent AI, an Antler-backed pre-seed startup. I specialise in LLM orchestration, agentic systems, and event-driven backend infrastructure — currently building the engine that helps Shopify merchants get discovered by ChatGPT, Claude, and Perplexity.
About Me
Hi, I'm Sanjay, a Backend & AI Engineer based in Hyderabad. I build the infrastructure that makes LLM-powered products work at scale — orchestration agents, event-driven batch systems, and multi-provider AI pipelines.
I'm currently the Founding Engineer at Comergent AI, an Antler-backed pre-seed startup building GEO/AEO infrastructure for Shopify merchants. I architected the entire backend from scratch — from a LangGraph multi-provider content agent running across 100+ stores concurrently, to an event-driven batch orchestrator that cut AI API calls by 73%.
Before that, I led the product engineering division at Qressy, shipping 12+ production Shopify apps serving 600+ active merchants. I also hold a LeetCode rating of 1552 (top 35% globally) and have 2,772 GitHub contributions in the last year.
Here are a few technologies I've been working with recently:
- LangGraph & LLM Agents
- Node.js & TypeScript
- Python & FastAPI
- AWS Lambda & EventBridge
- MCP (Model Context Protocol)
- MySQL, Redis & MongoDB
Where I’ve Worked
Founding Engineer @ Comergent AI
December 2025 - Present
- Reduced AI batch API calls by 73% by replacing naive polling with self-scheduling EventBridge rules using exponential backoff (2ⁿ mins, capped at 120), with provider-agnostic abstraction supporting Claude, OpenAI, and Gemini batch APIs
- Scaled LLM content generation to 100+ Shopify stores concurrently by building a LangGraph multi-provider orchestration agent (Python, FastAPI, asyncio) with dual-provider structured output abstraction — OpenAI uses native Structured Outputs, Anthropic uses tool_use with recursive schema patching, both returning identical Pydantic objects
- Built GEO visibility report engine running 12 parallel LLM calls per report across Claude, GPT-4o, Gemini, and Perplexity with a custom citation scorer normalizing patterns into a 0–100 GEO score
- Published npm MCP package (@qressy/qressy-meta-ads-mcp) exposing 40+ Meta Ads tools to Claude Desktop and Cursor via JSON-RPC 2.0 stdio-to-HTTP bridge with dynamic tool discovery from remote Lambda — full MCP protocol under 200 lines of TypeScript
- Owned end-to-end engineering: architecture decisions, production deployments, code reviews, and customer technical escalations across a platform supporting $1M+ GMV
Tech Stack
AI & LLM
Backend
AWS & DevOps
Databases
Shopify
Languages
Some Things I’ve Built
Featured Project
Comergent Batch Orchestrator
Event-driven AI batch job lifecycle manager supporting Claude, OpenAI, and Gemini batch APIs. Replaced naive polling with self-scheduling EventBridge rules using exponential backoff (2ⁿ mins, capped at 120), achieving a 73% reduction in API calls. Encrypted API keys in scheduler payloads, auto-deletes rules on job completion. Fully provider-agnostic abstraction under 200 lines of TypeScript.
- Node.js
- AWS Lambda
- EventBridge Scheduler
- Anthropic Claude
- OpenAI
- Google Gemini
![Comergent Batch Orchestrator]()
Featured Project
Comergent Content Agent
Multi-store LLM orchestration agent scaling content generation to 100+ Shopify stores concurrently. Dual-provider structured output abstraction — OpenAI uses native Structured Outputs, Anthropic uses tool_use with recursive schema patching — both returning identical Pydantic objects. Includes AI-detection evaluation loop, merchant alerting, title deduplication, and semaphore-based concurrency control.
- Python
- LangGraph
- FastAPI
- Anthropic Claude
- OpenAI
- Pinecone
- asyncio
![Comergent Content Agent]()
What's Next?
Get In Touch
I'm currently open to new opportunities and exciting projects! Whether you have a job opportunity, want to collaborate on a project, or just want to say hi, feel free to reach out. I'll try my best to get back to you!