Skip to content
Nnaveed
Résumé
Portrait of Shaik Mohammed Naveed

Software EngineerBackend & AI Systems
Hospet, India · IST

Shaik Mohammed Naveed

building MCP tool servers, agent gym environments, FastAPI services and evaluation pipelines

Python backend engineer at Turing. I build FastAPI services, MCP tool servers and the platform that runs GPT, Claude and Gemini through identical tasks, then checks the database to see which agent got it right.

See selected work
Résumé

Technologies I work with:

(01)About

I'm a backend engineer at Turing. I work on the platform where AI agent tasks get written, run against gym environments with MCP tools, and checked in the database.

Before that I automated cloud inventory as an intern at Zscaler. I graduated in Computer Science from BITM in 2024.

Naveed

Currently · Turing (Remote)Previously · Zscaler

Pull requests
520+

merged into a monorepo of FastAPI gym services

Years
2+

building Python backends on Google Cloud Platform

Weekly
1,500+

Colab notebooks validated by a workflow I built

Manual QA
90%

less manual QA effort on notebook delivery

Now

Building agent gym services and review tooling for AI agent tasks, with model batches running as Cloud Run jobs.

  • FastAPI
  • MCP
  • Cloud Run
  • PostgreSQL
Based in

Hospet, India

Remote · IST

IST · UTC+5:30
Indian PatentNo. 202441013223 AReal-time bus tracking(jump to Recognition)
IEEE ICDCECE 2024Co-authorIndexed in IEEE Xplore(jump to Recognition)
(02)Latest

A running log of recent work. Click an entry to open the full write-up.

9 entries · 23 Feb 2024 → 2024 – now

  6. a86b198Oct 2024MilestoneJoined TuringSoftware Engineer, Backend & AI Systems. Remote.
(03)Experience

Current role first. Open any card for the details.

AI agent evaluation

  • Built the backend for a multi-model benchmark that routes GPT, Claude and Gemini through identical tasks in isolated databases, scored by automated SQL verifiers.
  • 520+ merged PRs to a monorepo of FastAPI gym services that clone GitHub, Google Drive, Slack and Teams, each with MCP tools and a per-session SQLite database.
  • Shipped the AutoQC review workflow, trajectory evaluation with task versioning, and a Run Task feature that runs model batches as Cloud Run jobs.
  • Implemented MCP over JSON-RPC so agents call tools and run queries inside user permission scopes.

Backend & APIs

  • Built multi-tenant REST APIs for chat, channels and scheduling with hierarchical RBAC.
  • Built a FastAPI, Celery and BigQuery workflow that validates 1,500+ notebooks weekly. Manual QA dropped 90% and team client rework fell from 65% to 15%.
  • Wrote a PR velocity tracker on a daily cron with Plotly dashboards over BigQuery for engineering leadership.
  • FastAPI
  • Celery
  • Redis
  • PostgreSQL
  • MCP
  • JSON-RPC
  • Cloud Run
  • BigQuery

(04)Selected work

Most of my Turing work is internal, so these are write-ups with diagrams. Open any card for the full story.

9 projects

Turing · Oct 2024 – Present
Python · FastAPI · Cloud Run

The gym verification service: the internal platform for authoring AI agent tasks, running them against gym environments with MCP tools, and checking in the database whether the agent did the right thing.

review workflow
AutoQC
Cloud Run jobs
Run Task
over JSON-RPC
MCP
Turing · Oct 2024 – Present
Python · FastAPI · SQL

Backend for a multi-model AI benchmarking framework that routes GPT, Claude and Gemini through identical tasks inside isolated database environments.

model families
3
automated verifiers
SQL
database per run
Isolated
Turing · Oct 2024 – Present
Python · FastAPI · SQLite

A monorepo of FastAPI gym services that clone real platform APIs (GitHub, Google Drive, Slack, Teams) for AI agents to work in.

merged PRs
520+
platforms cloned
4
per session
SQLite
Turing · Oct 2024 – Present
FastAPI · Celery · BigQuery

An automated workflow that validates 1,500+ Colab notebooks weekly and converts them to standard JSON.

  • FastAPI
  • Celery
  • BigQuery
College · Patent filed 23 Feb 2024
React.js · Firebase · GPS

A React.js and Firebase bus tracker with geofenced route-deviation alerts, built on the low-latency GPS architecture in my Indian patent for real-time bus tracking and arrival prediction.

202441013223 A
Patent
low latency
GPS
deviation alerts
Geofence
(05)Stack

The rings turn slowly. Hover a tool to see where I've used it.Lime tools are the ones I use every day.
Pythonmain language

  • Languages

    • Python
    • SQL
    • TypeScript
    • Bash

  • Backend

    • FastAPI
    • Flask
    • Celery
    • SQLAlchemy
    • REST APIs
    • Async processing
    • pytest

  • Databases

    • PostgreSQL
    • SQLite
    • Redis
    • BigQuery

  • Cloud & DevOps

    • Google Cloud Platform
    • Cloud Run
    • Docker
    • Docker Compose
    • Git
    • Linux

  • AI & protocols

    • Model Context Protocol
    • JSON-RPC
    • OpenEnv
    • LangChain
    • LLM integration (OpenAI, Anthropic, Google)

  • Frontend & data

    • React
    • TypeScript
    • pandas
    • Plotly

Daily drivers

(06)Recognition

  • Patent2024

    Real-Time Bus Tracking and Arrival Prediction

    202441013223 A · filed 23 Feb 2024

  • Publication2024

    Home Automation using Wi-Fi: ESP32-based remote control

    ICDCECE 2024 · IEEE Xplore

  • CompetitionBITES

    Winner, BITES Robothon

    Autonomous robot design

  • Hackathon50+ teams

    First place, full-stack hackathon

    Out of 50+ teams

Certifications

  • Google AI EssentialsCoursera · 2024
  • Artificial Intelligence A-Z 2025Udemy · 2025
(07)Contact

Hiring for backend or AI infrastructure, or have something worth building? Send a note, or use the .

Open to backend & AI infrastructure roles

Hospet, India

 

naveedmohammed687@gmail.com
POST/api/contact

Send a message

What's this about?

Up to 4,000 characters. Press Command Enter to send.

Goes straight to my inbox. Only used to reply to you.