Abdur Rahim Sheikh

Software Engineer · Dhaka, Bangladesh · abi.rahim.sheikh@gmail.com

I am a CSE graduate from the University of Rajshahi with a deep passion for CS fundamentals, algorithmic optimization, and backend architecture. Currently, I specialize in building robust Python-based services that bridge the gap between complex data processing and AI. My approach to engineering is simple: write maintainable code, solve the hardest problems first, and never stop learning.

Work Experience

Software Engineer

MyMedicalHub International

Developing Backend AI services focusing on high-integrity data management and medical assistance.

  • Async Infrastructure: Transitioned legacy Stable Diffusion systems to a modular FastAPI + ComfyUI pipeline. Increased throughput from 5 to 30+ parallel requests/sec.
  • Legacy Refactoring: Refactored 3 years of technical debt into a maintainable codebase, significantly increasing development velocity for the core company backend project.
  • Docker & DevOps: Containerized microservices to ensure environment parity and streamlined deployment.
  • Collaboration: Working closely with clinical teams and stakeholders to translate complex medical requirements into technical logic.
August 2023 - Present

AI Research Assistant

MyMedicalHub International (Remote/Tampa, Florida)

Conducted deep research into medical AI while completing my undergraduate degree.

  • Built and custom-tuned models for object detection and classification.
  • Expertise in creating and annotating high-fidelity datasets for medical imaging.
  • Explored hyperparameter tuning for YOLO architectures to improve clinical detection accuracy.
October 2021 - December 2022

Problem Solving & Competitive Programming

Codeforces Expert

Max Rating: 1744 (Handle: ABIR)

Solved 770+ problems. This experience honed my ability to analyze time complexity and space complexity under pressure—skills I bring to every backend task.

LeetCode Consistency

900+ Problems Solved with a 500+ day continuous streak. I treat coding as a daily discipline.

Other Platforms & Impact:

  • LightOJ (184 Solves): Mastered the power of algorithms in worst-case scenarios.
  • UVA (76 Solves): Focused on extracting core problems from complex context-rich descriptions.
  • VJudge (379 Solves): Diverse practice across multiple global judges.

Detailed Projects

Clinical Rule Parser

Django Recursive Algorithms jQuery

Built a recursive parsing engine to evaluate 2,000+ medical variables. This project required deep application of tree-traversal logic to ensure real-time validation of complex medical conditions. View Demo

Medical Action Detection (Research)

YOLOv8 LSTM Temporal Flow

Classified 24 categories of human movement. Achieved 97% accuracy by applying median thresholding on Temporal Flow. This involved hand-annotating clinical data and building a custom detection pipeline.

QuantumMind Architecture

FastAPI Docker Ollama PostgreSQL

A containerized system capable of orchestrating multiple LLM-based apps. Integrated custom authentication and a task-queue mechanism for heavy generative workloads.

Technical Stack

Languages

Python, C++, SQL, JavaScript

Backend

FastAPI, Django, Redis, WebSockets, REST APIs

Tools

Docker, Git, PostgreSQL, Linux, Poetry/UV

AI/ML

PyTorch, LangChain, FAISS, Stable Diffusion, ComfyUI

Core Fundamentals

Advanced Data Structures, OOP, System Design, Unit Testing, Agile/Scrum

Key Achievements

  • Academic Excellence: Ranked 4th in the department (B.Sc. CSE) with a CGPA of 3.62/4.00.
  • Samsung R&D Coding Contest: 25th Place nationally among top-tier engineers.
  • ICT Division Job Festival 2023: 2nd Runner Up for "Journey through Bengal" game architecture.
  • Certified Docker Expert: Completed comprehensive certification covering local scripting to AWS deployment.