Hello, I'm
Python Backend Developer
I build backend applications and REST APIs with Python, FastAPI, and Flask. My work includes authentication, database-driven applications, API design, file handling, background processing, and backend testing using technologies such as MongoDB, MySQL, Redis, Docker, and pytest.
I am currently working as a Software Engineering Trainee Intern at FamilyShell LLP, where I work on the Health Records module of a production family management application. My work includes investigating production backend issues, improving existing architecture, and working with FastAPI, MongoDB, Redis, Docker, Cloudflare R2, and REST APIs.
Alongside my internship work, I build AI-integrated backend systems. My Investment Portfolio Tracker uses FastAPI, LangChain, LangGraph, RAG, and the Model Context Protocol (MCP), with pytest-based testing.
Production-style FastAPI backend application featuring investment portfolio management, JWT authentication, LangChain-powered RAG, LangGraph AI orchestration, Model Context Protocol (STDIO and Streamable HTTP), and automated pytest testing.
Tech Stack: FastAPI, LangChain, LangGraph, ChromaDB, SQLite, Docker, JWT, MCP, pytest.
Full-stack accessibility bug reporting platform featuring a feature-based FastAPI backend, MongoDB persistence, Redis-powered OTP authentication, JWT access and refresh tokens, Argon2 password hashing, role-based access control, and community bug verification.
Tech Stack: Python, FastAPI, MongoDB, PyMongo, Redis, JWT, Argon2, Flutter Web, Riverpod, Dio.
Production-style Flask REST API with MySQL persistence, layered architecture, secure JWT authentication, input validation, and deployment on Render.
Tech Stack: Flask, MySQL, JWT, bcrypt, Gunicorn, Docker.
May 2026 – Present
Some of the most valuable engineering lessons I have learned came from solving problems where the obvious fix was not the correct architectural solution.
Designed an identity and authorization flow for an MCP-enabled AI assistant without exposing internal database identifiers to the LLM. Implemented Personal Access Tokens and separated external resource references from internal database identity.
Investigated expired attachment URLs in the FamilyShell Health Records module and redesigned the storage-access flow around object paths and short-lived presigned URLs while preserving compatibility with existing production data.
Identified an architectural issue where reminders used the creator's timezone instead of the health-record subject's timezone. Redesigned timezone resolution and UTC scheduling around the actual subject of the reminder.