Samrand Hassan
English

Technical manager · Researcher

Samrand Hassan Leads Elryan’s store and marketplace engineering, and researches production machine learning behavior.

Over a decade of architecting and shipping resilient production systems across Kurdistan, China, and global remote engineering teams: spanning mission critical campus infrastructure, hardened enterprise ERP security, high scale e-commerce marketplaces, real time recommender pipelines, and sovereign language AI for Kurdish.

Samrand Mahmood Hassan
Now
Technical Manager at Elryan
Researching
Recommender reproducibility, and Kurdia — language technology for every Kurdish dialect
Studied
M.Sc. Software Engineering, Nankai University
Based in
Changsha, China · from the Kurdistan Region of Iraq
  • 10+Years architecting production systems
  • 27+Enterprise, campus, and commerce platforms shipped
  • 201Controlled experiments on ML production stability
  • 6Countries spanning industry & doctoral research

01Work

Selected work

All 39 projects

Platforms, integrations, and research I have designed or led: each broken down into the core problem, the architectural design, and what changed in production.

  1. 2025 – present

    Elryan commerce platform

    Frontend and middleware for a multi-system e-commerce platform and its vendor marketplace: storefront, apps, APIs and the integrations between them.
    ElryanPlatform · Live
  2. 2025 – present

    NetSuite ↔ vendor platform bridge

    A real-time bridge between Oracle NetSuite and the in-house vendor platforms, so products, stock and fulfilment stay in step without manual reconciliation.
    ElryanIntegration · Live
  3. 2026 – present

    Kurdia: Kurdish language technology

    Multi-dialect Kurdish language technology for standard hardware, covering dialect identification, spelling and grammar checking, optical character recognition (OCR), speech recognition, speech synthesis and a compact local language model.
    Personal researchResearch · In progress
  4. 2025 – 2026

    Silent ranking changes under dependency upgrades

    A controlled study examining the effect of NumPy, SciPy, pandas and PyTorch upgrades on recommender system outputs across 201 cases. I identify four causal mechanisms and observe no degradation in recommendation quality.
    Personal researchResearch · Open source
  5. 2025 – present

    Merchant & marketplace APIs

    The APIs vendors integrate against: catalogue sync into Magento, orders and stock, and rule-based commission profiles.
    ElryanIntegration · Live
  6. 2025 – 2026

    Open search & recommendation platform

    A fully open-source design for search, recommendations and an “Ask AI” assistant on top of Magento, Vue Storefront and Strapi.
    ElryanPlatform · Design stage
  7. 2025 – present

    Mitmana

    A live AI presentation product I built. Users can generate a deck from a prompt, start from Mitmana’s template directory or pick a layout from a marketplace of user-shared designs.
    PersonalProduct · Live
  8. 2026

    claude-switch

    Run work and personal Claude — Claude Code and Claude Desktop — side by side, fully isolated, with every sign-in routed to the right account.
    PersonalOpen source

02Research

Research

Publications and research

I focus on the production realities of machine learning: real time feature computation, recommender stability across library upgrades, and sovereign language AI for low resource, multi dialect languages.

Research interests

  • Recommender systems
  • Reproducibility of ML pipelines
  • SQL + ML feature computation
  • Scalable ML systems
  • Privacy-aware federated learning
  • Graph neural networks
  • Kurdish NLP across dialects
  • AI agents and the Model Context Protocol

0000-0001-6694-159X

  1. Privacy-Aware Compression for Federated Learning: A Comprehensive Survey of Mechanisms, Methods, and Trade-offs

    Samrand Hassan, Aras Aram, Ngin Ahmed

    International Journal of Latest Research in Engineering and Technology (IJLRET) 2025 Journal article

  2. Silent ranking changes under dependency upgrades in MovieLens recommender pipelines

    Silent Ranking Changes Under Dependency Upgrades in MovieLens Recommender Pipelines

    Samrand M. Hassan

    2026 Manuscript In preparation

  3. Optimization Techniques for SQL+ML Queries: A Performance Analysis of Real-Time Feature Computation in OpenMLDB

    Mashkhal A. Sidiq, Aras A. Salih, Samrand M. Hassan

    International Journal of Database Management Systems (IJDMS) 2025 Journal article

03Writing

Writing

All writing

  • A different list is not a worse listResearch note · What happens to a recommender when you upgrade NumPy: 201 controlled cases, four mechanisms, and no measurable loss of quality.

04Talks

Talks & media

See all

  • Cryptocurrency: the future of money? Talk How cryptocurrencies work, what problems they claim to solve, and what they might change about money.
  • Quantum computing Talk Qubits, superposition and entanglement, and why some problems become easier on a quantum computer.

05CV

Career

Full CV

From writing raw HTML at twelve to architecting mission critical platforms and leading engineering teams.

  1. 2025 – present Technical Manager and Engineering Team LeadManahil El Ryan LTD · Elryan
  2. 2022 – 2024 Senior Software Engineercodbel
  3. 2021 – 2022 IT Admin Officer (Software Developer)University of Kurdistan Hewlêr
  4. 2018 – 2021 Software Engineer, then Senior Software EngineerGaliawa Group
  5. 2019 – 2020 Senior Software EngineerJiasaz LLC for IT Services
  6. 2017 – 2018 Software DeveloperTEDxNishtiman
  7. 2022 – 2024 Master of Science in Software EngineeringNankai University
  8. 2014 – 2018 B.Sc. Software and Informatics EngineeringSalahaddin University–Erbil

Get in touch

Research collaboration, a talk, or a hard engineering problem — write to me.