Mohammed Khalid Shaik

$ ls projects/

Systems I built, what constrained them, and what changed as a result.

Enterprise Document Conversion API

Backend Engineer · Takeda ·2025-07–2026-07

An eight-endpoint async service that took per-file processing from over fifteen minutes to under thirty seconds for about a thousand users.

  • −97% processing time
  • ~1,000 active users
  • 100 concurrent sessions
  • 0 cross-session leaks
  • Python
  • FastAPI
  • asyncio
  • pytest
  • OpenAPI

Managed Analytics Platform

Application Engineer · Integrated Analytics Solutions ·2023-05–2025-03

A reconciliation service that drove isolated SAS Studio pods from database state, cutting secure provisioning from days to minutes.

  • days → minutes provisioning
  • ~1,000 concurrent users
  • 99.9% availability
  • 4 client contracts
  • Python
  • React
  • Material UI
  • Kubernetes
  • Docker
  • APScheduler

Promotion Approval Workflow

Data Engineer · GameStop ·2022-09–2023-05

An Airflow-orchestrated approval pipeline with database-backed tracking that cut team effort 80% and about $200K a month in leakage.

  • −80% team effort
  • ~$200K monthly saving
  • Python
  • Apache Airflow
  • SQL
  • REST APIs

Production LLM Agent Runtime

Application Engineer · Integrated Analytics Solutions ·2026-06–present

Trust and safety mechanics for an agent platform: canonical-hash verification, YAML write-safety, and a CI check that blocks stub implementations.

  • 71 endpoints shipped (team)
  • 5,900+ tests (team)
  • 100+ model providers (team)
  • < 4 months time to ship (team)
  • Python
  • FastAPI
  • WebSocket
  • PostgreSQL
  • pgvector
  • pytest