Available for opportunities

Hi, I'm Ashutosh Kumar Yadav

24, he/him

Backend engineer from India, building scalable systems and AI-driven automation at Razorpay.

Resume Contact Me
hey! 👋
Ashutosh Kumar Yadav - Backend Software Engineer at Razorpay
about me.

I'm a Backend Engineer at Razorpay, engineering high-frequency financial systems that handle 600+ monthly incorporations and intelligent AI onboarding flows.

Specializing in distributed architectures and AI-driven automation, I've optimized critical services for 3x performance gains and built resilient microservices using Golang and Java. I build MCP servers for LLM tool integration, design agentic AI workflows with LangGraph, and leverage RAG pipelines for intelligent data retrieval. Currently building NeuroVest—an AI-powered stock analysis platform using RAG, sentiment analysis, and predictive models.

When I'm not debugging distributed systems, you'll find me analyzing F1 telemetry 🏎️, playing football ⚽, or exploring new places 🌍.

skills.

Languages

Golang Java C++ JavaScript Python

Backend & Architecture

Microservices Distributed Systems gRPC Domain-Driven Design SDK Development REST APIs

Databases & Queues

MySQL Redis DynamoDB Kafka

Infrastructure & DevOps

Docker Kubernetes Helm Charts Spinnaker CI/CD Git

AI & LLMs

Prompt Engineering LLM Automation Agentic AI RAG Pipelines MCP Servers LangGraph OpenAI API Google Vision

Observability & Testing

Grafana Prometheus Logging & Metrics Alerts & Runbooks Unit Testing Integration Testing E2E Testing
experience.
Jul 2024 – Present

Software Development Engineer

Razorpay
  • Built Rize Incorporation Self-Serve Platform from 0→1, scaling to 600+ incorporations/month, 1100+ payments/month, reducing onboarding TAT from 48 hours to 2.5 minutes
  • Designed Conversational Onboarding using LangGraph, Supervisor & Domain Agents, GPT-4o, Redis state management, and cyclic/acyclic agent flows—enabling 250+ successful onboardings with highly accurate structured data extraction
  • Contributed 8+ reusable components to the Modular Onboarding Framework, adopted across 8+ applications to standardize UIs, validations, data models, and platform logic
  • Re-architected User-Service (~2000 RPM) with 40+ API migrations to Go, identifier-based proxy middleware, goroutine-based parallelization—achieving 3× performance improvement
  • Built document-verification fraud checks using OpenAI (GPT-4.1/4o/3.5) + Google Vision (95% accuracy) and automated workflows via MCP servers, RAG, agentic AI—resolved 80+ security tickets in 4 days
  • Built async lead-generation pipelines using goroutines + queue delays (5s→10s→20s) with cron retries; collaborated with FE/SF teams to resolve P0/P1 onboarding issues
  • Fixed email/phone update flows and invitation logic, improving experience for 600+ users/month and reducing 250+ monthly support tickets
  • Added Grafana dashboards, alerts, runbooks, and comprehensive logging/metrics with strong unit, integration, and E2E test coverage
Sep 2023 – Jul 2024

Product Development Engineer – Intern

Razorpay
  • Sole backend owner of Rize microservice; integrated Marketplace, launched D2C+ community, onboarded 2 communities from Excel to backend
  • Refactored service architecture using Domain-Driven Design with 82% test coverage
  • Automated merchant onboarding using Google Vision + OpenAI with 92% accuracy
  • Set up stage deployment pipelines using Helm on Kubernetes; fixed Docker image/version issues, resolved P0 incidents
  • Built a scalable logging framework that reduced operational overhead and improved debugging velocity
  • Fixed critical bugs in dedupe logic, Redis flows, cron jobs, and core APIs—improving system reliability
  • Wrote robust UTs, integration tests, and E2E tests for all changes; safely executed production DB migrations
projects.
Featured • In Development

NeuroVest

AI-Powered Stock Analysis Platform for the Indian Market — a "Glass Box AI" with 7 Deterministic Engines that compute scenarios, risk scores, and price zones. The LLM is a renderer, not a decision-maker. Built on a Hybrid-RAG architecture where AI is sandwiched between layers of hard mathematical verification.

7-Engine Core: Signal · Price · Risk · Scenario · Relative Strength · Backtesting · Sentiment
RAG Pipeline: ChromaDB vector search · GPT-4 synthesis · narrative guardrails
3-Tier DB: MySQL (18 tables) · ChromaDB (2 collections) · Redis (12+ key patterns)
4D Security: Rate limiting (device/session/IP/user) · JWT key rotation · anti-hallucination
Python FastAPI Next.js Azure OpenAI RAG ChromaDB Redis MySQL Docker Sentiment Analysis LangGraph
GitHub

Go MCP Weather Server

Model Context Protocol (MCP) server built in Go, providing real-time weather data through LLM tool discovery endpoints. Supports HTTP, SSE, and stdio transports for seamless integration with AI coding assistants like Cursor. Features clean architecture, TOML configuration, health checks, and graceful shutdown.

Go MCP Protocol OpenWeather API SSE Docker Cursor IDE
GitHub

ShopMart-X

Scalable E-Commerce Engine handling complex product lifecycles and secure order processing. Implements role-based access control (RBAC), optimized MySQL schema for high-volume transactions, and efficient inventory management.

Node.js MySQL EJS Tailwind
GitHub

CampusHive

Secure Content Management System implementing JWT-based stateless authentication and Bcrypt encryption. Designed for high-concurrency read operations with scalable MongoDB indexing strategies.

React MongoDB Node.js JWT
GitHub Live Demo

Sorting Visualizer

Algorithm Visualization Engine rendering real-time complexity analysis of O(n log n) sorting algorithms. Optimized for 60fps performance using efficient DOM manipulation and state management.

React JavaScript CSS
GitHub Live Demo

Face Recognition App

High-performance Computer Vision system using MediaPipe and OpenCV pipelines. optimized for low-latency inference (60fps) and robust real-time face tracking in varying lighting conditions.

Python MediaPipe OpenCV
GitHub

Mini API Projects

Collection of mini projects consuming various APIs including Weather API, RestCountries, and NameFake for data generation.

JavaScript REST APIs HTML/CSS
View All on GitHub

40+ Full-Stack Projects

Collection of CRUD applications, Canvas tools, and various full-stack projects built to explore different technologies and patterns.

React Node.js MongoDB Canvas API
View All on GitHub
education.
🎓
B.E. Computer Science Chitkara University, Punjab, India
2020 – 2024 • CGPA 9.7
achievements.
💻

DSA

Solved 200+ problems on LeetCode, 100+ across GFG and other platforms

🎃

Hacktoberfest 2021

4/4 PRs accepted in one week during Hacktoberfest

✨

Spark AR Hackathons

Created AR filters for social applications

🚀

Building at Scale

Built platform handling 600+ incorporations and 1100+ payments monthly at Razorpay

⚡

Performance Engineering

Achieved 3× performance improvement on User-Service architecture

🤖

AI Integration Pioneer

Resolved 80+ security tickets in 4 days using AI-powered document verification

contact me :-)

I'm currently open to new opportunities and interesting projects. Feel free to reach out!

Email akyadav3996@gmail.com
Phone +91 8447344852