Satish Kumar

Python Backend Developer

I build APIs, integrations and automations that hold up in production — FastAPI, Flask & Django services, webhooks, and AI-powered backends.

About

Backend & API Integration Specialist

Software Developer — Backend & API Integration with 5+ years of experience building Python backends, REST APIs and third-party integrations. Currently at Apeejay Stya Group (VJAL) in Noida, India, where I work on API integrations and automation; earlier at Trident Group India and OPPO India.

My day-to-day is the unglamorous work that keeps products running: plugging payment, CRM and messaging APIs into one reliable layer, making webhooks survive retries and failures, and shipping FastAPI / Flask / Django services on AWS. Lately that extends to AI — LLM support agents, RAG services and MCP tool servers, several of which are open below.

Core stack

  • Python
  • FastAPI
  • Flask
  • Django
  • REST APIs
  • API Integrations
  • Webhooks
  • AWS · EC2 / RDS / Lambda
  • n8n
  • Xano
  • MySQL
  • Redis
  • Docker
  • LLM Apps · RAG
  • MCP Servers
  • Git

Experience

  • Jul 2024 — Present

    Software Developer — Backend & API Integration

    Apeejay Stya Group (VJAL) · Noida, India

    Backend services and API integrations in Python — including a FastAPI + OpenAI document Q&A assistant used on real documents, plus workflow automation with n8n and Xano.

  • Earlier

    Software Developer

    Trident Group India · OPPO India

    Backend development and API work across product and internal systems.

  • 2019

    Intern

    AT&T

    Early industry experience in software development.

Projects

Open-source work from my GitHub — AI services, integration tooling and production backend patterns. Every card links to the live repository.

AI Support Agent

LLM-powered ticket triage, drafting & routing — FastAPI

An AI agent for support inboxes: it classifies incoming tickets with an LLM, drafts a reply in the right tone, and routes each ticket to the correct queue — with webhook notifications so the team sees new and escalated tickets instantly. Built to sit in front of a real helpdesk, not replace it: drafts are suggestions a human approves.

  • FastAPI
  • LLM APIs
  • Webhooks
  • Python

MCP Tool Server

Backend tools exposed for LLM tool-calling — Model Context Protocol

A Model Context Protocol server that gives LLM assistants real hands: database lookups, API calls and utility tools exposed over MCP, so a model can fetch live data and take actions instead of guessing. Clean tool schemas, typed inputs and safe defaults — the pattern I use when connecting LLMs to production backends.

  • MCP
  • Python
  • LLM Tool-Calling
  • FastAPI

RAG API Service

Retrieval-augmented generation API — FastAPI + LangChain

Ask questions over your own documents and get answers grounded in them. The service ingests files, chunks and embeds them, stores the vectors, and exposes a Q&A endpoint that retrieves the relevant passages before the LLM answers — so responses cite the source material instead of hallucinating.

  • FastAPI
  • LangChain
  • Embeddings
  • Vector Search

API Integration Hub

Payments, CRM & messaging APIs behind one unified layer

The pattern from my day job, as clean open code: multiple third-party APIs — payments, CRM, messaging, email — wrapped in a single FastAPI client layer with consistent auth handling, automatic retries, and rate-limit backoff. Adding a new provider means writing one adapter, not re-plumbing the whole app.

  • FastAPI
  • REST
  • Retries & Backoff
  • Rate Limiting
View on GitHub Integrations

Webhook Relay Service

HMAC-verified webhook intake with retries & delivery logs

Webhooks fail quietly — endpoints go down, signatures get checked wrong, events get lost. This service receives webhooks, verifies HMAC signatures, fans deliveries out to subscriber endpoints, retries failures with exponential backoff, and keeps a full delivery log so nothing disappears silently.

  • Python
  • HMAC Verification
  • Backoff Retries
  • Delivery Logs
View on GitHub Integrations

FastAPI Production Starter

Production-grade backend: JWT auth, SQLAlchemy, Redis, Docker

The FastAPI setup I reach for on real projects, packaged as a starter: JWT authentication, SQLAlchemy with MySQL, Redis caching, a clean project layout, tests, and Docker — so a new service starts at production-shaped instead of tutorial-shaped.

  • FastAPI
  • JWT
  • SQLAlchemy · MySQL
  • Redis
  • Docker

FastAPI DOCX Bot

Document Q&A assistant — FastAPI + OpenAI, built at work

A real work project: upload a DOCX document and ask questions about it in plain language — the assistant answers from the document's own content using OpenAI models behind a FastAPI service. The kind of internal tool that saves a team hours of Ctrl+F.

  • FastAPI
  • OpenAI
  • DOCX Parsing
  • Python
View on GitHub AI & LLM · Backend

Let’s talk backend work

Have a backend, API or integration project in mind? The fastest way to reach me is email — or find me on LinkedIn and GitHub.