I'm a Backend & AI Platform Engineer building production-grade backend systems and AI platforms across B2B commerce, SaaS, and developer tools.

• 🤖 Currently building WizPilot, an agentic AI assistant for B2B commerce, backed by WizCommerce's shared Almanac AI platform.

• 🧠 Working on multi-agent systems, LLM infrastructure, MCP, RAG, tool calling, AI guardrails, and agent orchestration using Python, LangGraph, LangChain, and FastAPI.

• ⚙️ Built an MCP platform with 200+ enterprise tools, multi-tenant agent execution, dynamic tool discovery, and tool-schema compression reducing LLM token overhead by ~90%.

• 📊 Built production RAG and NL-to-SQL systems using Qdrant, OpenAI embeddings, Voyage reranking, Elasticsearch, and BigQuery.

• 🚀 Strong backend foundation in distributed systems, microservices, API optimization, databases, system design, and performance engineering.

• 🏓 Represented state-level U16 in table tennis — discipline and competitiveness fuel my engineering journey.

WizCommerce (Oritur Technologies Pvt Ltd)
Sep 2025 — Present | Bengaluru, India

• Co-architected Almanac, WizCommerce's shared AI platform powering WizPilot, WizStudio, Ella, and Workflow Engine.
• Built WizPilot, an agentic AI assistant integrating WizOrder with CRM, catalog, cart, analytics, and order management for natural-language B2B workflows.
• Designed a multi-tenant, database-configured agent execution layer with per-tenant graph caching, PostgreSQL conversation state, and dynamic MCP/subagent registration.
• Built an MCP platform exposing 200+ enterprise tools with tenant-aware routing, validation, guardrails, and secure execution.
• Implemented dynamic MCP tool compression, reducing LLM schema token overhead by ~90%.
• Built multi-agent orchestration using LangGraph and LangChain, with domain-specialist agents and structured invocation APIs.
• Implemented multi-provider LLM routing through LiteLLM across OpenAI, Anthropic, and Google models.
• Built production RAG NL-to-SQL analytics pipelines using OpenAI embeddings, Qdrant, Voyage reranking, Elasticsearch, and BigQuery.
• Built a tenant-scoped Knowledge Base with hybrid BM25 + dense retrieval and RRF.
• Implemented SSE streaming, conversation memory, progressive-disclosure agent skills, HITL controls, and AI guardrails.
• Developed high-performance FastAPI microservices, reducing critical API latency by 80%, system errors by 32%, and increasing throughput by 35%.
• Built a parallelized migration pipeline processing 1M+ records at 1,000+ docs/sec between PostgreSQL and Elasticsearch.

Tech: Python, FastAPI, LangGraph, LangChain, MCP, LiteLLM, PostgreSQL, Qdrant, Elasticsearch, BigQuery, Redis, AWS, Docker, Kubernetes

Fynd (Shopsense Retail Technologies Ltd)
Feb 2025 — Aug 2025 | Mumbai, India

• Built backend services for Fynix AI, an AI-powered coding assistant supporting IntelliJ and VS Code.
• Improved codebase indexing by 90%+ through embedding optimization and vector database tuning.
• Designed modular ETL and ingestion pipelines for GitHub, GitLab, Azure DevOps, Quip, Google Docs, and BigQuery.
• Engineered large-scale ingestion pipelines for Stack Overflow and public GitHub datasets.
• Built RAG workflows and LLM-powered features using LangChain, LangGraph, OpenAI, and Claude.

Tech: Python, Java, Kotlin, LangChain, LangGraph, OpenAI, Claude, Vector Databases

Growder Technovations & Ompax Lifestyle Pvt. Ltd.
Mar 2023 — May 2024 | Surat, India

• Engineered backend architectures across two early-stage B2B platforms.
• Designed scalable Order Management and User Management systems with JWT authentication and RBAC.
• Improved backend throughput by 70–80% through performance optimization and architectural improvements.
• Built reusable integrations with Shiprocket, Shopify, payment gateways, logistics, and third-party services.
• Automated warehouse, logistics, and inventory workflows using event-driven microservices.

Tech: Java, Spring Boot, PostgreSQL, MySQL, Redis, RabbitMQ, Elasticsearch, AWS

Agentic AI assistant for B2B commerce built on WizCommerce's shared Almanac AI platform.

Highlights:

• Multi-agent orchestration
• MCP with 200+ enterprise tools
• Multi-tenant agent execution
• RAG & NL-to-SQL
• Qdrant + Elasticsearch
• LiteLLM multi-provider routing
• SSE streaming
• Conversation memory
• AI guardrails & HITL
• Structured agent invocation

Shared AI infrastructure powering multiple WizCommerce products.

Designed to provide reusable:

• Agent orchestration
• LLM infrastructure
• MCP tool execution
• Tool discovery & compression
• Multi-tenant isolation
• Agent memory
• RAG pipelines
• Structured AI outputs
• Observability & evaluations

AI-powered coding assistant with IDE integrations.

Built backend services supporting:

• Codebase indexing
• Embeddings
• Vector search
• RAG workflows
• GitHub / GitLab ingestion
• Stack Overflow datasets
• LLM-powered code assistance

• ⚡ 80% reduction in critical API latency
• 📉 32% reduction in backend errors
• 🚀 35% increase in system throughput
• 🧠 ~90% reduction in LLM tool-schema token overhead
• 🔧 200+ enterprise AI tools through MCP
• 📦 1M+ records migrated at 1,000+ docs/sec
• 💾 70% reduction in repetitive database reads
• 🔍 90%+ improvement in codebase indexing at Fynix AI
• 🏗️ Multiple products built from 0 → 1

🎓 Master of Computer Applications — Machine Learning & Artificial Intelligence
Amity University, Noida | 2025

🎓 Data Science & Machine Learning — Scaler DSML
2025

🎓 Full Stack Web Development — Masai School
Bengaluru | 2023

🎓 Bachelor of Business Administration — BBA
Birla Institute of Technology, Mesra | 2020

• 🧠 Solved 300+ DSA problems across competitive programming platforms
• 🚀 Built and scaled multiple 0 → 1 products
• 🤖 Production experience with Agentic AI, LLMs, RAG, MCP, and Multi-Agent Systems
• ⚙️ Strong foundation in Distributed Systems, System Design, LLD, HLD, and API Design
• 🏓 Represented state-level U16 in table tennis