About Mahdi Molavi

Mahdi Molavi October 4, 2026 5 min read

Mahdi Molavi is a senior .NET backend engineer and a tensor decomposition researcher based in Izmir, Türkiye. For more than ten years he has built high-traffic systems with C#, ASP.NET Core, SQL Server and Redis, and in his M.Sc. thesis at Tarbiat Modares University he developed tc-SVD, a tensor decomposition built on a cosine-transform product.

This page collects the facts in one place: short and long bios, the numbers behind the work, the research, and where to find him. Everything here matches his CV and the projects described on this site.

Bios

One line

Mahdi Molavi is a senior .NET backend engineer who builds high-traffic systems and researches tensor decompositions for big data.

Short (about 50 words)

Mahdi Molavi is a senior backend engineer with more than ten years of experience in C#, ASP.NET Core, SQL Server and Redis. He has built assessment platforms for tens of thousands of concurrent users and leads .NET microservices work. He holds an M.Sc. in Computer Science, and in his thesis he developed the tc-SVD tensor decomposition.

Longer

Mahdi Molavi is a senior .NET backend engineer and tensor researcher based in Izmir, Türkiye. Since October 2023 he has been a senior backend engineer at NovaTech, where he architected an adaptive testing platform based on item response theory that serves about 89,000 students, leads .NET 6/8 microservices development and mentors the backend team. Before that, at Gozine2, he designed and deployed a Redis Sentinel high-availability setup that sustained 45,000 to 70,000 concurrent users at exam peaks, tuned SQL Server and hot APIs with Locust load tests, and founded a four-person data science team. On a 2024 contract for Noraco he built a PDF-to-LaTeX pipeline with LLM tagging for a bank of more than 5,000 exam questions.

He studied applied mathematics at Urmia University and computer science at Tarbiat Modares University. His thesis, Big Data Learning with Tensor Representation, supervised by Dr. Mansoor Rezghi, introduced tc-SVD and used it for classification, clustering, brain imaging data and streaming updates. He has also been a teaching assistant for Advanced Machine Learning and Data Mining.

At a glance

  • Role: Senior Backend Engineer at NovaTech (remote), since October 2023.
  • Experience: more than ten years with .NET: C#, ASP.NET Core, EF Core, SQL Server, Redis, RabbitMQ, Docker.
  • Scale: an adaptive testing platform for about 89,000 students; a Redis Sentinel setup for 45,000 to 70,000 concurrent users at exam peaks; a question bank of 5,000+ items built with an LLM-assisted pipeline.
  • Research: tensor decompositions (t-SVD and tc-SVD) for classification, clustering and incremental learning on big data.
  • Education: M.Sc. in Computer Science, Tarbiat Modares University, 2016–2019 (GPA 18.8/20); B.Sc. in Applied Mathematics, Urmia University, 2011–2015.
  • Teaching: teaching assistant for Advanced Machine Learning and Data Mining (2018, 2020–2021).
  • Awards: rank 138 of 50,000+ in Quera competitive programming; top 5 of 2,000+ in the HamCode C# competition (2021); 3rd of 5,000+ in HamCode JavaScript (2021); finalist in the Fanavard machine learning competition (2018).
  • Languages: Azerbaijani Turkish and Persian (native), Turkish and English (professional).
  • Location: Izmir, Türkiye; works remotely.

Career

  • 2023–present, NovaTech: senior backend engineer; IRT-based adaptive testing platform, .NET microservices lead.
  • 2024, Noraco (contract): PDF-to-LaTeX and LLM tagging pipeline, FastAPI services and a React review console.
  • 2021–2023, Gozine2: senior backend engineer; Redis Sentinel high availability, SQL Server and API tuning, data science team lead. Earlier in 2021, a full-stack role that redesigned reporting and introduced structured logging.
  • 2020–2021, Refah Bank exchange house (freelance): a back-office system that replaced spreadsheet-based FX remittance accounting.
  • Since December 2018, freelance and side projects: Marketo, a multi-store marketplace with React Native apps (2018–2020); a topic map and wiki for a banking research institute’s encyclopedia and a web-based BPMS (2019–2020); a crypto and forex market-data pipeline on TimescaleDB as part of a team (2024); a delivery and field-service platform (2025–2026); Hamsayeh, a building-management SaaS, and a multi-tenant microservice platform in progress (2026).
  • 2019, OmidanHozoor: stabilized React and React Native apps and built versioned, documented REST APIs.
  • 2016–2018, Withem: .NET developer on a private social network for actors.

Some of this work is written up as case studies under Projects.

Research

His research asks how to learn from large multiway data without flattening it into vectors first. The work builds on the t-product and t-SVD of Kilmer and Martin and replaces their periodic boundary with a reflective one; the result, tc-SVD, runs on the discrete cosine transform. The tensor guides explain the ideas in short form, the thesis summary covers the results, and the MATLAB code is open source at github.com/mahdipc/tcSVD.

Open source

  • tcSVD: tc-SVD and t-SVD in MATLAB, with tensor-based classification, clustering and incremental updates.
  • Dima.Finance: a financial management backend in ASP.NET Core with Clean Architecture, DDD and CQRS.
  • Numerical analysis codes: Mathematica notebooks that reproduce experiments from ten papers on integral equations, wavelet methods and Runge–Kutta schemes, on GitHub.

Writing

He writes about tensor methods, big data and backend engineering. Start with Tensor Notes 1: Stop Flattening Your Data or the short guide to t-SVD.

Profiles and contact

For websites, backend and performance work, search visibility or tensor methods, see Work with me.

Frequently asked questions

Who is Mahdi Molavi?

Mahdi Molavi is a senior .NET backend engineer with more than ten years of experience building high-traffic systems in C#, ASP.NET Core, SQL Server and Redis, and a tensor decomposition researcher. He holds an M.Sc. in Computer Science from Tarbiat Modares University and lives in Izmir, Türkiye.

What is Mahdi Molavi known for?

In industry, for high-traffic .NET systems: an adaptive testing platform serving about 89,000 students and a Redis Sentinel setup that sustained 45,000 to 70,000 concurrent users at exam peaks. In research, for tc-SVD, a tensor decomposition built on a cosine-transform product.

What is tc-SVD?

tc-SVD is the tensor decomposition Mahdi Molavi developed in his M.Sc. thesis. It replaces the periodic boundary hidden in t-SVD with a reflective one, so the computation runs on the discrete cosine transform and stays in real arithmetic.

Where is Mahdi Molavi based?

He lives in Izmir, Türkiye, and works remotely with teams in other countries.

How can I contact Mahdi Molavi?

By email at mhdmlv93@gmail.com or through LinkedIn at linkedin.com/in/mahdimolavi. Project enquiries are described on the Work with me page.

Share
Back to the home page